No full-text available
To read the full-text of this research,
you can request a copy directly from the authors.
An Mr Brain Images Classifier via Principal Component Analysis and Kernel Support Vector Machine
Abstract
Automated and accurate classification of MR brain images is extremely important for medical analysis and interpretation. Over the last decade numerous methods have already been proposed. In this paper, we presented a novel method to classify a given MR brain image as normal or abnormal. The proposed method first employed wavelet transform to extract features from images, followed by applying principle component analysis (PCA) to reduce the dimensions of features. The reduced features were submitted to a kernel support vector machine (KSVM). The strategy of K-fold stratified cross validation was used to enhance generalization of KSVM. We chose seven common brain diseases (glioma, meningioma, Alzheimer's disease, Alzheimer's disease plus visual agnosia, Pick's disease, sarcoma, and Huntington's disease) as abnormal brains, and collected 160 MR brain images (20 normal and 140 abnormal) from Harvard Medical School website. We performed our proposed methods with four different kernels, and found that the GRB kernel achieves the highest classification accuracy as 99.38%. The LIN, HPOL, and IPOL kernel achieves 95%, 96.88%, and 98.12%, respectively. We also compared our method to those from literatures in the last decade, and the results showed our DWT+PCA+KSVM with GRB kernel still achieved the best accurate classification results. The averaged processing time for a 256 × 256 size image on a laptop of P4 IBM with 3 GHz processor and 2 GB RAM is 0.0448 s. From the experimental data, our method was effective and rapid. It could be applied to the field of MR brain image classification and can assist the doctors to diagnose where a patient is normal or abnormal to certain degrees.
... Because it allows for image analysis at multiple levels of resolution, the wavelet transform is an excellent technique for extracting characteristics from Magnetic Resonance (MR) brain images. As a result of its multi-resolution analytical capacity [1]. For this purpose, researchers have submitted several studies, all of which have yielded positive results. ...
... Multifilter construction approaches are already being developed to make use of them [12][13][14]. The objective of this multiplicity is to achieve the characteristics described below [1]. ...
... DWT provides local signal time-frequency information with cascaded high-pass and low-pass filter banks to hierarchically retrieved features, enabling the extraction of the most relevant features of varied pathways and dimensions [19]. The concept of Multiwavelet was born out of the generalization of scalar wavelets [1,2]. Multiple wavelets and scalar functions are used instead of a single scalar one. ...
A brain tumor is one of the most common and deadly diseases in the current era, which requires early and reliable detection techniques. The diagnosis of brain tumors depends on the analysis of Magnetic Resonance Imaging (MRI). It is a technology that provides the highest and most realistic images of the anatomical structures of the human body, particularly the brain. It includes information that can be used in place of a clinical diagnosis. It is a reliable method and does Citation: Al-Jawher WAM, Al-taee SHA (2022) Precise Classification of Brain Magnetic Resonance Imaging (MRIs) using Gray Wolf Optimization (GWO) J Brain Neursci 6: 021.
... Because it allows for image analysis at multiple levels of resolution, the wavelet transform is an excellent technique for extracting characteristics from Magnetic Resonance (MR) brain images. As a result of its multi-resolution analytical capacity [1]. For this purpose, researchers have submitted several studies, all of which have yielded positive results. ...
... Multifilter construction approaches are already being developed to make use of them [12][13][14]. The objective of this multiplicity is to achieve the characteristics described below [1]. ...
... DWT provides local signal time-frequency information with cascaded high-pass and low-pass filter banks to hierarchically retrieved features, enabling the extraction of the most relevant features of varied pathways and dimensions [19]. The concept of Multiwavelet was born out of the generalization of scalar wavelets [1,2]. Multiple wavelets and scalar functions are used instead of a single scalar one. ...
A brain tumor is one of the most common and deadly diseases in the current era, which requires early and reliable detection techniques. The diagnosis of brain tumors depends on the analysis of Magnetic Resonance Imaging (MRI). It is a technology that provides the highest and most realistic images of the anatomical structures of the human body, particularly the brain. It includes information that can be used in place of a clinical diagnosis. It is a reliable method and does Citation: Al-Jawher WAM, Al-taee SHA (2022) Precise Classification of Brain Magnetic Resonance Imaging (MRIs) using Gray Wolf Optimization (GWO) J Brain Neursci 6: 021.
... In Step 1, we first used Eq. (1) to delineate the KEDF landscape of an input brain tumor image I ∈ R H×W acquired from an open-source dataset 33 . Then we constructed the kernel k r ′ ; r based on its spatial characteristics in 2-dimensional Euclidean. ...
... Moreover, magnet www.nature.com/scientificreports/ Step 1, the mapping of the FLAIR image 33 constructs the KEDF landscape, followed by forming the reciprocal distance kernel (RDK). In Step 2, element-wise products of 2D-FFTs of the image and kernel are calculated and then inversed for PEDF estimation. ...
... The CPU time for this small dataset ranges between 0.06 to 0.09 s. We acquired these images from an open-source dataset 33 www.nature.com/scientificreports/ like several low-energy impurities filled below the Fermi surface; they obstruct the electron transmission along the surface and thus cover the exertion of intrinsic material properties. ...
We demonstrate that isomorphically mapping gray-level medical image matrices onto energy spaces underlying the framework of fast data density functional transform (fDDFT) can achieve the unsupervised recognition of lesion morphology. By introducing the architecture of geometric deep learning and metrics of graph neural networks, gridized density functionals of the fDDFT establish an unsupervised feature-aware mechanism with global convolutional kernels to extract the most likely lesion boundaries and produce lesion segmentation. An AutoEncoder-assisted module reduces the computational complexity from O N 3 to O N log N , thus efficiently speeding up global convolutional operations. We validate their performance utilizing various open-access datasets and discuss limitations. The inference time of each object in large three-dimensional datasets is 1.76 s on average. The proposed gridized density functionals have activation capability synergized with gradient ascent operations, hence can be modularized and embedded in pipelines of modern deep neural networks. Algorithms of geometric stability and similarity convergence also raise the accuracy of unsupervised recognition and segmentation of lesion images. Their performance achieves the standard requirement for conventional deep neural networks; the median dice score is higher than 0.75. The experiment shows that the synergy of fDDFT and a naïve neural network improves the training and inference time by 58% and 51%, respectively, and the dice score raises to 0.9415. This advantage facilitates fast computational modeling in interdisciplinary applications and clinical investigation.
... The factual technique for finding the surfaces to consider the spatial association of the pixels is Gray Level Co-Occurrence Matrix abbreviated as GLCM. [4] The target of GLCM is to recognize the outside of a picture by assessing the pixel in pair with explicit qualities and determined spatial relationship. GLCM helps in estimating the factual extricated parameters. ...
... Every segment of the pixels which comes under the mask is grouped together to represent a single pixel value and this grouping is based on average values of all pixels inside mask. (4) Where N is the number of pixels inside the window and Xi represents each pixel. ...
... Support Vector Machine Classifier is used to classify classes based on a hyper-plane which is similar to a line cutting a plane into two parts and hence also known as a discriminative classifier. [4] Each class is on either side of the hyper plane. Thus SVM is used when the input data is to be classified among these classes only. ...
This paper proposes a methodology in which detection, extraction and classification of brain tumour is done with the help of a patient’s MRI image. Processing of medical images is currently a huge emerging issue and it has attracted lots of research all over the globe. Several techniques have been developed so far to process the images efficiently and extract out their important features. The paper describes certain strategies including some noise removal filters, grayscaling, segmentation along with morphological operations which are needed to extract out the features from the input image and SVM classifier for classification purpose.
... The model uses labeled data to learn the structure of the problem and then uses the unlabeled data to fine-tune its predictions [1]. Neural networks (NN) [2], support vector machine [3], K-nearest neighbour, and linear regression are widely used supervised learning algorithms. An artificial neural network (ANN) is a computational model which is inspired by the human brain. ...
... Twodimensional (2D) DWT is a versatile signal-processing tool formulated by Daubechies and Mallet in the late 1980s. DWT splits a signal into low-frequency and high-frequency parts, called decomposition [3]. DWT decomposition of an image into low resolution, i.e., approximation coefficients (LL subband) and high resolution, i.e., detail coefficients (the horizontal HL, the vertical LH, and the diagonal HH subbands). ...
... The results obtained in this paper using the KMC+DWT+SCGBPNN algorithm and KMC+DWT+BRBPNN algorithm are compared with other existing literature, i.e., DWT+HC+BPNN [9], DWT+HC+SVM [9], DWT+SOM [16], DWT+linear kernel SVM [16], DWT+PCA+Linear kernel SVM [3], DWT+PCA+KSVM(HPOL) [3], and DWT+PCA+ANN [2]. Table 6 shows the compilation of various image classification algorithms, the number of images in the dataset, and classification accuracy. ...
Brain tumors and brain cancer can have varying
prognoses based on factors such as tumor type and location,
patient age, and overall health. Early detection of brain tumors
is crucial for saving lives, and magnetic resonance imaging
(MRI) is a fast and safe method for detecting these tumors.
However, there is a need for non-invasive and efficient brain
image analysis techniques to aid in tumor detection. This paper
presents a technique for brain image analysis that utilizes kmeans
clustering to segment images and wavelet transform to
extract features. A total of 22 texture features are extracted and
normalized before being submitted to a classifier. The dataset of
253 brain images is classified as tumor or non-tumor using
BPNN classifiers. The scaled conjugate gradient-based
backpropagation neural network (SCG-BPNN) and bayesian
regularization-based backpropagation neural network (BRBPNN)
are used for training. The BR-BPNN training algorithm
outperforms the SCG-BPNN algorithm with an accuracy of
97.23%, although it takes longer to train. The proposed work is
compared with other existing research and found to perform
better in terms of classification accuracy.
... Other advantages of quadratic kernels include having fine mathematical tractability and direct geometric representation. 19,[22][23][24][25] K -Nearest neighbors. K-NN is a popular classifier used commonly because of its simplicity, straightforwardness, and high efficiency despite noisy data. ...
... The kerosene is loaded through three loading pumps (PUMP_4,5,6). Each loading pump is fed by kerosene through a suction valve (MOV_20, 22,24) and discharged through a discharge valve (MOV_21, 23,25). The Auto/ Manual Push buttons are used to select the control mode of the loading pump. ...
The paper introduces a Programmable Logic Controller (PLC) / Human Machine Interface (HMI) system incorporated along with machine learning (ML) classifiers. Although some other studies have incorporated ML techniques to predict and control petroleum product terminals in terms of concentration, the proposed framework incorporates Add On Instruction (AOI) programming, PLC, and ML methods to automatically monitor petroleum products terminals. The framework adds an AOI in programming to achieve maximum usage of processor capabilities. Moreover, it uses AOI for programming in cooperation with the ladder diagram (LD). This leads to simplifying the LD graphical programming language, reducing the time of scanning, and making facilitate troubleshooting. The AOI is merged with ML to automate tank level detection and maintain good operational conditions and consequently protect these expensive essential assets. The introduced framework consists of three stages. The first stage is the PLC programming phase where the PLC is created using Add-On instructions. Next, HMI graphic displays are drawn and linked to the PLC tags in the following stage. During the third stage, the actual process readings are applied to the system based on ML algorithms to test its functionality. The proposed system results indicates a reduction in the LD number, highest program size, and maximum time of scanning. The results indicate that the AOI can help to trace the program more easily in fault situations. Besides, additional program instructions could reduce processor memory, system construction costs, and upgrade projects.
... Other advantages of quadratic kernels include having fine mathematical tractability and direct geometric representation. 19,[22][23][24][25] K -Nearest neighbors. K-NN is a popular classifier used commonly because of its simplicity, straightforwardness, and high efficiency despite noisy data. ...
... The kerosene is loaded through three loading pumps (PUMP_4,5,6). Each loading pump is fed by kerosene through a suction valve (MOV_20, 22,24) and discharged through a discharge valve (MOV_21, 23,25). The Auto/ Manual Push buttons are used to select the control mode of the loading pump. ...
This paper proposes a PLC (Programmable Logic Controller)/HMI (Human Machine Interface) system for tracking oil products refineries. The proposed pipeline includes AOI (Add on Instruction) programming, and PLC to automatically display petroleum products terminal. It provides an AOI in programming to obtain the most utilization of processor capabilities. Besides, it makes use of AOI for programming in cooperation with a ladder logic program. This results in simplifying the ladder program, decreasing scan time, and making troubleshooting easier. The proposed system is constructed in two stages. First is the PLC controller programming stage. In the second stage, the HMI graphic presentations are drawn and connected to the PLC tags. The proposed system outcomes confirmed that the number of logic ladders, maximum program size, and maximum scan time has decreased. The outcomes imply that the AOI can assist in tracing the program more without difficulty in faults situations. Besides, it adds extra program commands in less processor memory, lowering system creation, and upgrade costs.
... Surgery is used to remove the aberrant brain cells, and subsequent treatment with medication is intensive. An abnormal development of tissue that may be seen in the brain is called a tumor [10][11][12][13][14]. Brain tumors are very rare, and unlike tumors in other regions of the body, they can only spread from one brain cell to another inside the brain. ...
The article describes the use of image processing in the search for the brain tumor. Tumors are a worldwide health crisis that may manifest in any part of the body. If a brain tumor is not diagnosed and treated early on, it may significantly reduce a patient's lifespan. Malignant and benign tumors of varied stages were discovered. This illness is now affecting a sizable population. Each year, more and more people are screened every day in an effort to discover diseases early. Manual screening is not only a time-consuming process, but it also raises the possibility of making mistakes. Some people could become even more distracted. Therefore, it is more ideal to use an AI-based system for the screening process than a human one. Specialists employ MRI (Magnetic Resonance Imaging) scan images to identify brain cancers; however, these images include noise that must be minimized in the first phases of processing if accurate findings are to be achieved. In this research, we provide a refined Discrete Wavelet Transform (DWT) filtering technique along the used of artificial Intelligence for this purpose. Image filtering and image segmentation are the system's two main components. Brain tumor MRI images from the Kaggle dataset will be used as test data for the filter.
... Despite the advancement of wavelet analysis, several wavelet types have become more and more well-known. The most important wavelet is the Harr wavelet, which is the simplest one and often the preferred wavelet in a lot of applications 45 . Equation (3) and Eq. ...
... Despite the advancement of wavelet analysis, several wavelet types have become more and more well-known. The most important wavelet is the Harr wavelet, which is the simplest one and often the preferred wavelet in a lot of applications 45 . Equation (3) and Eq. ...
Brain tumors are among the most fatal and devastating diseases, and they often result in a significant reduction in life expectancy. The devising of treatment plans that can extend the lives of affected individuals hinges on an accurate diagnosis of these tumors. Identifying and analyzing large volumes of magnetic resonance imaging (MRI) data manually proves to be both challenging and time-consuming. As a result, there exists a pressing need for a reliable machine-learning approach to accurately diagnose brain tumors, and numerous methods have already been proposed over the last decade. In this paper, a novel, comprehensive approach is proposed for identifying and classifying a given MR brain image as abnormal. Three common brain diseases, namely glioma, meningioma, and pituitary tumor, are chosen as abnormal brains, and the Figshare MRI brain image dataset was collected from the Kaggle and IEEE websites. The proposed method is initiated by employing 1st-order statistics, 2nd-order statistics, and higher-order transformed (DWT) feature extraction to extract features from images. Then missing data is addressed and handled using KNNImputer, followed by the application of the ExtratreesClassifier and PCA feature selection methods to identify the most relevant features and reduce the dimensions of these features. Subsequently, the reduced features are submitted to seven machine learning models, namely RF, GB, CB, SVM, LGBM, DT, and LR. The strategy of k-fold cross-validation is utilized to enhance the performance of those models. Finally, the models are evaluated using XAI approaches, which ensure transparent decision-making processes and provide insights into the model’s predictions. Remarkably, our approach achieves the highest accuracy, precision, recall, F1 score, MCC, Kappa, AUC-ROC, and R2, as well as the lowest loss, among the seven models evaluated, proving its effectiveness and applicability in multiple analytic applications relying on publicly available datasets.
... Following the evaluation process, an impressive accuracy of 99% was achieved. Similarly, the authors of [80] [81] [82] [8] [83] used PCA for feature reduction. It is a statistical technique employed in data analysis to simplify complex datasets. ...
This manuscript provides a detailed review of state-of-the-art techniques for identifying brain tumors using magnetic resonance imaging (MRI) analysis, focusing on automatic brain tumor detection (ABTD). The study is driven by the critical need to improve the accuracy and effectiveness of tumor identification in clinical settings, where traditional methods can be prone to manual error and extensive. The review covers five key stages: data collection, preprocessing, segmentation, feature extraction, and classification. Techniques are systematically categorized and evaluated based on features, performance metrics, and clinical applicability. A structured framework is proposed for assessing ABTD methods, offering a comprehensive comparison of pros, cons, processing speed, and detection accuracy. The review highlights the effectiveness of various techniques, with the combination of LuNet and Deep Convolutional Neural Networks (DCNN) obtained 99.7% accuracy. Areas for improvement in current methodologies are identified, and future research directions are suggested, supported by quantitative analysis and statistical evaluation. This work aims to advance the development of more reliable and practical ABTD techniques for use in real-world clinical environments.
... This has been made possible by the fact that ACO adopts filters that enable it select only those features that are most relevant in improving the accuracy and robustness of these models for the purpose of early diagnosis. Furthermore, ACO can easily scale medical data with numerous dimensions such as image medical data, genetic information and patients' histories, which are applied in predictive models of lung cancer [36][37]. This shows that ACO can be effective in these applications to be placed as highly efficient optimization tool in the sector of medical diagnostics. ...
Lung cancer is still ranked among the major killers and current techniques used in the diagnosis of this condition do not assist in early detection. In this paper, the research seeks to work towards resolving the early detection of lung cancer using ACO and more specifically, the hybrid learning models. The ACO optimization algorithm based on foraging ants is used to select the features contributing to model simplification and increased performance. The approach incorporates ACO with DNN to enhance the system's diagnostic performance. The performance of the proposed approach has been evaluated and analyzed for different scenarios using experimental results, it has been shown that the accuracy of the DNN model is relatively increased and achieved up to 0. 91 to 0. 97 indicating that the application of ACO in breast cancer detection has the ability of minimizing false positives and thus improving the chances of early diagnosis. This line of work is a major advance toward a more accurate diagnosis of lung cancer and can be extrapolated to the field of more targeted health care.
... It is possible to make images with great contrast by adjusting the radio frequency and gradient pulse, for example. There are two types of brain tumors: benign and malignant [3]. Non-cancerous tumors are benign, and cancerous tumors, called malignant tumors, are more likely to develop as a result of cancer in any region of the body, not just the brain. ...
Medicine is one of the fields where the advancement of computer science is making significant progress. Some diseases require an immediate diagnosis in order to improve patient outcomes. The usage of computers in medicine improves precision and accelerates data processing and diagnosis. In order to categorize biological images, hybrid machine learning, a combination of various deep learning approaches, was utilized, and a meta-heuristic algorithm was provided in this research. In addition, two different medical datasets were introduced, one covering the magnetic resonance imaging (MRI) of brain tumors and the other dealing with chest X-rays (CXRs) of COVID-19. These datasets were introduced to the combination network that contained deep learning techniques, which were based on a convolutional neural network (CNN) or autoencoder, to extract features and combine them with the next step of the meta-heuristic algorithm in order to select optimal features using the particle swarm optimization (PSO) algorithm. This combination sought to reduce the dimensionality of the datasets while maintaining the original performance of the data. This is considered an innovative method and ensures highly accurate classification results across various medical datasets. Several classifiers were employed to predict the diseases. The COVID-19 dataset found that the highest accuracy was 99.76% using the combination of CNN-PSO-SVM. In comparison, the brain tumor dataset obtained 99.51% accuracy, the highest accuracy derived using the combination method of autoencoder-PSO-KNN.
... But this method can differentiate only glioblastoma, sarcoma and metastatic bronchogenic carcinoma tumors. Zhang and Wu [5] proposed a classification method by using kernel support vector machine (KSVM). Wavelet transforms as DWT was used to extract features from images and reduced the dimensions of those features by PCA method. ...
Magnetic resonance imaging (MRI) of the brain is essential for measuring and visualizing the brains anatomical structure, analyzing brain abnormalities, delineating pathological regions, surgical planning, and image-guided interventions. Image processing techniques are applied to MRI images for identification, detection, and classification of brain diseases. In this paper, an improved feature extraction method is proposed, which can extract the features of brain MRI images for detecting brain tumors as benign and malignant. In the proposed method Log-Polar Transformation (LPT) based feature extraction method is developed to extract the features from brain MRI images which are used as detection factor of brain abnormalities. Kernel support vector machine (KSVM) is applied to the extracted features as image classification tools for the tested brain MRI images. Extensive experiments are simulated over different orientation and conversion of scales in T-1 and T-2 weighted MRI images. A comparative analysis among the proposed method with other promising method is performed, where it is clearly shown the improvement of the proposed method.
... A brain tumor is described as an abnormal growth of tissue [10][11][12][13][14]. Brain tumors, unlike malignancies in other regions of the body, can only spread within the brain. ...
Brain cancer classification is a difficult task due to the variety and complexity of tumors shown in magnetic resonance imaging (MRI) pictures. This research presents two neural network approaches for categorizing MRI brain images. The proposed neural network method consists of three steps: feature extraction, dimensionality reduction, and classification. First, we extracted features from MRI images using discrete wavelet transformation (DWT). In this second stage, we reduce the salient features of MRIs using principal component analysis (PCA). For the classification step, two supervised machine learning classifiers have been developed. Artificial neural networks are used by both classifiers; however, the second one employs back propagation (BPN) while the first one uses feed-forward (FF-ANN). Using the classifiers, MRI brain images of the subjects were classified as normal or abnormal. Artificial neural networks have numerous applications, including function approximation, feature extraction, optimization, and classification (ANNs). They are specifically intended to enhance photos, distinguish and categorize items, separate and register objects, and extract features. Among these, object and picture recognition is the most important for complex processing tasks such as classifying brain tumors. Radial basis function (RBF), cellular, multi-layer perceptron (MLP), hop field, and pulse-coupled neural networks have all been used in image segmentation. These networks can be categorized as feed-forward (associative) or feedback (auto-associative).
... It interprets this information to create our perception of the world around us. (ii) Motor Control: The brain controls voluntary and involuntary movements of the body [7]. It sends signals to muscles and other organs through the nervous system to enable movement, from simple actions like picking up an object to complex tasks like playing a musical instrument [8]. (iii) Homeostasis: The brain helps maintain the body's internal balance (homeostasis) by regulating functions like body temperature, blood pressure, and hormone production [9][10][11]. ...
Pathological Brain Detection (PBD) is a crucial field aimed at identifying and diagnosing structural or functional abnormalities in the brain related to neurological, neurodegenerative, or psychiatric disorders. This detection process typically employs medical imaging technologies such as MRI, CT, or PET scans, along with neurological evaluations, blood tests, and other diagnostic tools. Early detection is essential for effective treatment, improving prognosis, and enhancing the quality of life for affected individuals. Each common computational method provides unique insights into brain pathologies. Despite significant advancements in technology and methodology, PBD faces challenges such as variability in brain anatomy, the complexity of disorders, data quality, standardization issues, interpretability of models, and ethical concerns. Addressing these challenges necessitates collaboration among researchers, clinicians, and policymakers to develop robust and ethical methods for improving the detection and diagnosis of brain disorders.
... For the past 40 decades, it is used as a significant tool in scientific research and medicine diagnose. MRI is the most suitable choice for investigating brain tumor as it is more sensitive than CT scan in identifying small tumors and it gives better visualization [4]. ...
4) 2: 730-738] © 2018 IJSRST | Volume 4 | Issue 2 | Print ABSTRACT A Brain tumor is very serious disease causing deaths of many individuals. The detection and classification system must be available so that it can be diagnosed at early stages. Detection of the brain tumor and its type in its early stage is essential for right treatment. So classification of brain tumor is very important. Tumor classification has been one of the most challenging tasks in clinical diagnosis. Different image processing techniques such as image segmentation, image enhancement and feature extraction are used for detection of the brain tumor in the MRI images of the cancer affected patients. Medical Image Processing is the fast growing and challenging field now days. Image processing and neural network techniques are used to improve the performance of detecting and classifying brain tumor in MRI images. The objective of this review paper is to presents a comprehensive overview for MRI brain tumor segmentation methods. In this paper, various segmentation techniques have been discussed. Comparative analysis of existing techniques has been done in brief.
... [17] presented a Support Vector Machine (SVM)-Active learning technique that allows the operator to tag chosen images that are closer to the SVM border to train the SVM. According to [53], when all of the feedback rounds have been completed, the SVM is trained to extract the top n imageries that are distant from the support vector machine border. [36] put forward a content-centric RS for retrieving images. ...
Content-Based Medical Image Retrieval is used to extract meaningful information from a large number of medical images. To extract image texture characteristics, a novel medical image retrieval technique is presented. Images from Computed Tomography scans are used as input images. At first, the respective image is affected by some noise disturbance. To mitigate this and to improve the contrast of the pixels in the source image, and then a fractional Hartley transform is applied to eradicate the noise variation as well as reduce the image distortion. About this, a better-filtered output is determined. Then the hybrid feature extraction technique is utilized to extract the desirable features. After this Modified Weight-Brownian Motion Monarch Butterfly Optimization approach is exploited to alleviate the unwanted features from the huge amount of features and select certain desirable features. Finally, the similarity between the selected features is measured to detect and classify the medical images. The proposed technique has been more advantageous to increase the accuracy level as well as gradually minimizing the error rate. The proposed algorithms are implemented using the image processing toolbox in MATLAB 2019 platform. The performances are analyzed by the proposed method in terms of precision (99%), recall (83%), F-measure (90%), and accuracy (98.85%).
... It is used here to locate the necessary signal information to classify our mammogram images. Thus, we used the 2D discrete wavelet transform, which led to four sub-bands Low-Low (LL), High-Low (HL), Low-High(LH), High-High(HH), with a three-level decomposition of our image [20]. The wavelet approximations at the first and second levels are represented by LL1, LL2 respectively. ...
Radiologists use mammogram images for the diagnosis of breast cancer. However, interpreting these images remains challenging depending on the type of breast, especially on dense breasts. Dense breasts may contain abnormal structures similar to normal breast tissue and could lead to a high rate of false positives and false positives negatives. We present an efficient computer-aided diagnostic system for detecting and classifying breast masses. After removing noise and artifacts from the images
using 2D median filtering, mathematical morphology and pectoral muscle removal by Hough's algorithm, the resulting image is used for breast mass segmentation using the watershed algorithm. After the segmentation, the system extracts several data by the wavelet transform and the co-occurrence matrix (GLCM) to finally lead to a classification as malign or benign mass via the Support Vector Machine (SVM) classifier. This method was applied to 48 Medio-Lateral Oblique (MLO) images from the image base (mini-MIAS). The algorithm showed a 87.5% classification rate, 92.59%
sensitivity, and a specificity of 93.94%.
... In our study, we have applied a linear and quadratic kernel SVM to classify the datasets. QSVM is a nonlinear kernel with excellent mathematical adaptability and direct geometric explanation [28] that may outperform LSVM. LSVM and QSVM kernel equations are shown in Eq. (1) and (2) [29]. ...
Breast cancer is one of the main causes of mortality for women around the world. Such mortality rate could be reduced if it is possible to diagnose breast cancer at the primary stage. It is hard to determine the causes of this disease that may lead to the development of breast cancer. But it is still important in predicting the probability of cancer. We can assess the likelihood of occurrence of breast cancer using machine learning algorithms and routine diagnosis data. Although a variety of patient information attributes are stored in cancer datasets not all of the attributes are important in predicting cancer. In such situations, feature selection approaches can be applied to keep the pertinent feature set. In this research, a comprehensive analysis of Machine Learning (ML) classification algorithms with and without feature selection on Wisconsin Breast Cancer Original (WBCO), Wisconsin Diagnosis Breast Cancer (WDBC), and Wisconsin Prognosis Breast Cancer (WPBC) datasets is performed for breast cancer prediction. We employed wrapper-based feature selection and three different classifiers Logistic Regression (LR), Linear Support Vector Machine (LSVM), and Quadratic Support Vector Machine (QSVM) for breast cancer prediction. Based on experimental results, it is shown that the LR classifier with feature selection performs significantly better with an accuracy of 97.1% and 83.5% on WBCO and WPBC datasets respectively. On WDBC datasets, the result reveals that the QSVM classifier without feature selection achieved an accuracy of 97.9% and these results outperform the existing methods.
... They achieved an accuracy of 96.51%. Zhang et al. [10] proposed an automatic method for classification of MRI brain images based kernel support vector machine (KSVM) and wavelet transform (WT) features with Principal Component Analysis (PCA) to reduce the size of features. Usman and Rajpoot [11] investigated wavelet texture features with random forest classifier to predict tumor labels as multiclass classification. ...
... 3 Brain tumor classification is a significant progressing investigation area. [5][6][7] Extensively two types of classification methods are such as supervised and unsupervised 8 strategies. Two major unsupervised methods are self-organization map (SOM) and fuzzy C-means (FCM) which cannot categorize difficult cancers as high-grade cancers. ...
The brain tumor classification is implemented through biopsy, which is not normally executed before classic mind surgery. Machine learning (ML) algorithms assist radiologists in tumor analysis, not including obtrusive evaluations. The conventional ML strategies need separate feature extraction to tumor detect thus it needs more computation time to perform classification. Deep learning (DL) based convolution neural networks (CNNs) have been focused on brain tumor detection. In this paper, the CNN algorithm is improved based on meta‐heuristics, which are used for pre‐trained systems for databases to categorize MRI brain tumor images. Pre‐trained DL, binary swallow swarm optimization (BSSO) is used for improving the weight and predispositions of the CNN algorithm. It is a block‐wise calibrating system which is dependent on transfer learning. The current technique is assessed over a publically accessible magnetic resonance imaging (MRI) brain tumor database containing three categories as glioma, meningioma, and pituitary by the most noteworthy rate among everyone brain tumor in medical training. The proposed strategy is assessed over T1‐weighted contrast‐enhanced MRI (CE‐MRI) benchmark data. To assess the execution, utilize the proposed strategies to the CE‐MRI dataset for tumor detection and in the general execution of the BSSO‐CNN model is estimated using the execution assessment measurements such as precision, sensitivity (recall), specificity, F1‐score, and accuracy. Exploratory outcomes demonstrated with the purpose of the proposed strategy higher when compared to other methods to all metrics.
... Traditional methods generally extract a series of features such as statistical features and texture features of images, and then, use artificial neural networks, random forests, and support vector machines for segmentation [4]. Traditional methods can be roughly divided into four categories, namely threshold-based segmentation methods [5,6], edge-based segmentation methods [7][8][9], and cluster-based segmentation methods [10][11][12] and region-based segmentation methods [13,14]. However, the pros and cons of extracting features in traditional methods will greatly affect the final results of the experiment. ...
In recent years, the fully convolutional network represented by Unet has been widely used in the field of medical image segmentation. However, due to the diversity of the shapes of lesions and the differences in the structures of different organs, the segmentation of lesions using only Unet structure cannot meet the requirements of accuracy and speed. Therefore, an improved Unet network for brain tumor segmentation is proposed. To reduce the number of parameters while extracting richer features and improving the accuracy of segmentation, this article introduces the inverted residuals block to replace the convolution module in the encoding and decoding stages of Unet to improve the calculation speed and accuracy; to better combine high-order semantic information with low-order semantic information, improve for the quality of detailed features in the training process, an improved Residuals Convolutional Block Attention Module is added between the encoder and the decoder. Combining the above two points of improvement, this article proposes an improved model based on Unet. Based on the Brats2019 dataset, an ablation experiment was performed on the proposed improved Unet model and compared with the TrUE-Net, ConResNet and OM-Net methods, and the Dice coefficient and Hausdorff distance were used as evaluation indicators to analyze the segmentation effect of the model. The experimental results show that the Dice coefficient of the improved Unet network model proposed in this article is 0.020–0.027 higher than other comparative models on average, and the Haushofer distance is reduced by 2.67–10.06.
... But this method can differentiate only glioblastoma, sarcoma and metastatic bronchogenic carcinoma tumors. Zhang and Wu [5] proposed a classification method by using kernel support vector machine (KSVM). Wavelet transforms as DWT was used to extract features from images and reduced the dimensions of those features by PCA method. ...
Magnetic resonance imaging (MRI) of the brain is essential for measuring and visualizing the brains anatomical structure, analyzing brain abnormalities, delineating pathological regions, surgical planning, and image-guided interventions. Image processing techniques are applied to MRI images for identification, detection, and classification of brain diseases. In this paper, an improved feature extraction method is proposed, which can extract the features of brain MRI images for detecting brain tumors as benign and malignant. In the proposed method Log-Polar Transformation (LPT) based feature extraction method is developed to extract the features from brain MRI images which are used as detection factor of brain abnormalities. Kernel support vector machine (KSVM) is applied to the extracted features as image classification tools for the tested brain MRI images. Extensive experiments are simulated over different orientation and conversion of scales in T-1 and T-2 weighted MRI images. A comparative analysis among the proposed method with other promising method is performed, where it is clearly shown the improvement of the proposed method.
... With the purpose of providing a full-resolution semantic prediction, a U-Net augments the normal CNN design by adding a comparable expanding path. The goal is to create segmentation images that highlight specific features and objects in the image [23]. ...
Signal and image processing is a part of biomedical science. In that, Biomedical image processing have a great value such as recognition, segmentation and classification for disease diagnosis. In one part of biomedical science, brain tumor classification is considered with Magnetic Resonance Images (MRI) images using state of art models. Initially, the Convolutional Neural Network (CNN), Fast Convolutional Neural Network (FCNN), U-Net and M-Net model was considered for classification. Further, the Hybrid Firefly Meta Optimization (HFMO) is proposed for the better prediction purpose. The proposed work has three phases like normalization with augmentation, deep attention segmentation and classification. In the first phase, data augmentation is applied to increase the scope of the dataset. In the second phase, a deep attention technique is applied to concentrate on hotspot in the image during segmentation. Finally, a hybrid firefly optimization is applied to enhance the performance of the model in convolution neural network by backtracking the process. The measuring parameters like Dice coefficient, Jaccard index, Positive Projected Value (PPV), True Positive Rate and False Positive Rate were evaluated. The comparative analysis of various state of art models with proposed classifier were demonstrated. Thus the proposed technique produces the training accuracy as 98.62%, testing accuracy as 95.31 % and 1 % of loss.
... Discreate Wavelet Transform is employed to extract multiple features from the filtered image. Fourier transformation has a serious disadvantage as it doesn't contain any information about time domain [14]. LL, LH, HH, and HL are the four sub-bands for each scale [15]. ...
Brain tumor is a deathly disease and it is indispensable to point out the tumor immediately. Detection of brain tumor from MRI with higher accuracy has become a major research region for the medical sector. In this paper, an automatic brain tumor classification procedure applying fuzzy C-means and artificial neural network is proposed which provides higher precision. For this proposed technique, inputted MRI images are resized and then sharpening filter is used for preprocessing. After that, fuzzy C-means cluster process is chosen for image segmentation. At the next step, discreate wavelet transform is utilized for feature extraction and then features quantity are reduced by principal component analysis. Furthermore, reduced features are taken to artificial neural network for brain tumor classification. An effective training function Levenberg–Marquardt is used for neural network. This proposed method provides 99.8% accuracy, 100% sensitivity and 99.59% specificity which is comparatively better than other existing detection techniques.KeywordsBrain tumorFuzzy C-meansDiscreate wavelet transformArtificial neural networkLevenberg–Marquardt function
... Therefore, many studies about systems to classify brain tumor types have been conducted. Zhang et al [1] extracted features with wavelet transform (WT) and reduced dimensions using principal component analysis (PCA), then classify the tumors with a support vector machine (SVM). Othman el al [2] proposed a probabilistic neural network with PCA to classify meningioma and glioma. ...
... Also, DWT was used with a naive Bayes classifier (NBC) to improve brain image categorization [27]. Most of the models employ DWT and PCA for feature extraction besides feature selection and are evaluated with different classifiers such as KNN, feed-forward backpropagation artificial neural network (FPANN) [28], forward neural network (FNN) with adaptive chaotic particle swarm optimization (ACPSO) [29] and scaled chaotic artificial bee colony (SCABC) [30], backpropagation neural network (BPNN) with scaled conjugate gradient (SCG) [31], kernel support vector machine (KSVM) with gaussian radial basis (GRB) kernel [32] and least squares support vector machine (LSSVM) [33]. A LSSVM was used with PCA in classifying the brain MR images that utilize Ripplet transform type for capturing the features [34]. ...
Neuroimaging, a part of medical imaging is playing a crucial role in the field of neuroscience and psychology for analysing the structure as well as functions of the internal nervous system. It is the most evolving and prominent method in medical science as it not only focuses on the structure, but also the functional aspects of the brain. It helps in diagnosing metabolic and neurodegenerative diseases. More precisely, magnetic resonance imaging is one of the cutting-edge technologies in the medical verdict for the past four decades in diagnosing Neuro disorders. In recent years, machine learning and other high-end reckoning tools are used universally in assisting the automated identification of neurodegenerative diseases such as Parkinson's disease (PD), Alzheimer's disease (AD) and so on. However, similarities exist in the functional part of the nervous system, making it difficult for the learning algorithms to classify the disease. The main goal of this article is to predict neurodegenerative diseases in an effective way using functional analysis of brain images. The proposed model for predicting neurodegenerative diseases based on brain image analysis has two main phases. The first phase processes the image by applying pre-processing techniques and extracts the features from the images using wavelet transform and histogram gradients for further processing. The second phase applies machine learning algorithms such as multilinear principal component analysis and regularized discriminant analysis for selecting the features and identifying the prediction labels respectively. The experimental analysis has been performed with various brain image datasets such as DS-66, DS-75, DS-160 and DS-255. The obtained results show the 100% accuracy on DS-66, DS-75, and DS-160 datasets and 99.58% accuracy on DS-255 and take a minimum computational overhead of 0.025s. Thus, the proposed model offers improved results than many other existing models. Also, the performance analysis using various quality metrics endorses the significance of the proposed model in diagnosing neurodegenerative disease.
... P. Kumar et al. [16] used four stage method for segmentation and classification in which principal component analysis (PCA) was utilized to reduce dimensionality and afterward SVM classifier was applied for segmentation and managed to achieve accuracy level of 94%. Similar method was also applied by Y. Zhang et al. [17] as he used kernel SVM and for enhancement K-fold stratified cross validation was applied. GUOLI SONG et al. [18] also did similar work but in addition applied GLCM for hybrid feature extraction and was able to attain 98.36% accuracy. ...
The segmentation and detection of brain pathologies in medical images is an indispensible step. This helps the radiologist to diagnose a variety of brain deformity and helps in the set up for a suitable treatment. Magnetic Resonance Imaging (MRI) plays a significant character in the research area of neuroscience. The proposed work is a study and probing of different classification techniques used for automated detection and segmentation of brain tumor from MRI in the field of machine learning. This paper try to present the feature extraction from raw MRI and fed the same to four classifier named as, Support Vector Machine (SVM), Decision Tree (DT), k-Nearest Neighbors (KNN), and Artificial Neural Network (ANN). This mechanism was done in various stages for Computer Aided Detection System. In the preliminary stage the pre-processing and post-processing of MR image enhancement is done. This was done as the processed image is more likely suitable for the analysis. Then the k-means clustering is used to sectioning the MRI by applied mean gray level method. In the subsequent stage, statistical feature analysis were done, the features were computed using Haralick’s equation for feature based on the Gray Level Co-occurrence Matrix. Feature chosen dependent on tumor region, location, periphery, and color from the sectioned image is then classified by applying the classification techniques. In the third stage SVM, DT, ANN, and KNN classifiers were used for diagnoses. The performances of the classifiers are tested and evaluated successfully.
This research presents a heart disease prediction system aimed at identifying individuals at risk based on their medical history. With the rising incidence of heart-related conditions, early diagnosis is essential for timely intervention. The system utilizes machine learning algorithms, specifically Logistic Regression and K-Nearest Neighbors (KNN), to classify patients as likely or unlikely to develop heart disease. Experimental results demonstrate improved prediction accuracy compared to traditional methods like Naïve Bayes. The proposed model not only enhances diagnostic precision but also contributes to cost-effective and efficient healthcare. The implementation is provided in .pynb format for practical usability.
Early detection of diseases is a crucial step towards patient healthcare and recovery, especially, for low survival rate diseases like brain cancer. In recent years, deep learning models such as convolutional neural networks (CNNs) have been widely utilized for medical image classification owing to their proficiency in extracting local features. However, accurate classification of complex medical images necessitates capturing local and global features. This paper focuses on the classification of the three most prominent kinds of brain tumors, namely, glioma, meningioma and pituitary tumor, with the proposed CNN based on the structure of GoogLeNet. The proposed model uses an attention-based inception module (ABIM) embedded with attention mechanism of spatial attention and efficient channel attention. Attention helps CNN focus on features with more distinguishing power and ignores irrelevant features while the inception approach aids in extracting multiscale local features. ABIM incorporates attention with inception to generate better feature representation. The model training and testing are done on a publicly accessible Figshare dataset with 3064 brain MRI samples. The dataset comprises 1426, 708, and 930 MRI samples of glioma, meningioma, and pituitary tumors, respectively. In a five-fold cross-validation experiment, the proposed model achieved an average accuracy of 97.97% for glioma, 94.76% for meningioma, and 99.22% for pituitary tumors. The proposed CNN model surpasses existing models with an overall accuracy of 97.62%, while utilizing fewer parameters than its base model, GoogLeNet. The model is also evaluated on datasets from the Kaggle and Harvard medical repositories, and showed promising results in distinguishing between different tumor classes, and normal brain MRIs. These results demonstrate that our model not only outperforms other existing models in terms of accuracy but also has comparatively lower computational complexity, providing a robust tool to support healthcare professionals.
Brain tumor diseases have had a considerable impact worldwide, affecting millions of individuals of different age groups, including both children and adults above 20 years old. Due to they are more needed in people’s lives, using the method based classifying brain tumors by machine learning schemes has become necessary. However, healthcare applications face challenges in identifying the most suitable classification-based metric, such as accuracy, due to the utilization of recent datasets. This study paper aims to provide a thorough evaluation of computational intelligence strategies used in tumor diagnosis. Several successful data mining techniques have been implemented, including wavelet analysis and spatial pixel modulation techniques. Furthermore, feature extraction and reduction techniques, such as the Grey Level Co-occurrence Matrix (GLCM), have been used to prepare the features for classification. Magnetic resonance imaging scan (MRI) is frequently utilized for the diagnosis of brain tumor diseases which is highly applied for classification-based machine learning, The review paper was focused on gliomas, meningiomas, and pituitary adenoma diseases. Technically, the usage of kernel principal component KPCA analysis with the proposed adaptive back propagation neural network scheme produced better performance-based classification metrics, (i.e:99.84%) for the accuracy metric. The aforementioned review articles have demonstrated that usage of the machine learning-based health care applications (brain diseases) classification widely assists the patient’s outcome and operations inside the hospitals. In summary, the paper has highlighted the importance of machine learning schemes for brain tumor detection and classification, and it also provided a comprehensive analysis and comparison of the state-of-the-art to show the methods such as ;(feature extraction, feature reduction), pros, cons, and the contributions for each of them. The paper's results are considered an advantageous starting point for future works.
Challenging, iterative, error-prone and time-consuming process it is to classify, segment, and detect the area of infection in MRI images of brain tumors. Moreover, to visualize and numerically quantify the properties of the structure of the abnormal human brain even with sophisticated Magnetic Resonance Imaging techniques requires advanced and expensive tools. MRI can better differentiate and clarify the neuronal architecture of the human brain compared to other imaging methodologies. In this study, a complete pipeline is proposed to classify abnormal structures in the human brain from MRI images that might be early signs of tumor formation. The proposed pipeline consists of noise reduction techniques, gray-level matrix (GLCM) extraction features, segmentation of DWT-based brain tumor areas reducing complexity, and Support Vector Machine (SVM) using the Radial Basis Function kernel (RBF) in ensemble with PNN for classification. SVM and PNN in combination provide a data-driven prediction model of the possible existence and location of a brain tumor in MRI images Experimental results achieved nearly 99% accuracy in identifying healthy and tumorous tissue based on structure from brain MRI images. The proposed method together with comparable accuracy is reasonably lightweight and fast compared to the other existing deep learning-based methods.Keywordspre-processingimage segmentationDWTGLCMPCAMRI tumor ClassificationFeature extractionSupport vector machineProbabilistic Neural Network
Non-invasive grading of brain tumors provides a valuable understanding of tumor growth that helps choose the proper treatment. In this paper, an online method with an innovative optimization approach as well as a new and fast tumor segmentation method is proposed for the fully automated grading of brain tumors in magnetic resonance (MR) images. First, the tumor is segmented based on two characteristics of the tumor appearance (intensity and edges information). Second, the features of the tumor region are extracted. Then, the online support vector machine with the kernel (OSVMK) by dynamic fuzzy rule-based optimization of the parameters is used for the grading of tumors. The performance evaluation of the proposed tumor segmentation method was performed by manual segmentation using similarity criteria. Also, tumor grading results compared the proposed online method, the conventional online method, and the batch SVM with the kernel (batch SVMK) in terms of accuracy, precision, recall, specificity, and execution times. The segmentation results show a good correlation between the tumor segmented by the proposed method and by experts manually. Also, the grading results based on the accuracy, precision, recall, and specificity, 95.20%, 97.87%, 96.48%, and 96.45%, respectively, indicate the acceptable performance of the proposed method. The execution times of the introduced online method are much less than the batch SVMK. The method demonstrates the potential of fully automated tumor grading to provide a non-invasive diagnosis in order to determine the treatment strategy for the disease. So the physicians, according to the tumor's grade, can match the treatment of the brain tumor to the patient's individual needs and thus make the best course of treatment for each patient.
Brain tumors can be detected and diagnosed early, potentially saving many lives. The Magnetic Resonance Image (MRI) technique is extensively used, less invasive, considered as the best image scanning technique in medical imaging for diagnosing various diseases. However, filtering and denoising MR Images remain a challenge. There is no direction in the literature regarding the proper technique for MR image processing. The present article proposes a new hybrid model for brain MR image processing that employs the Otsu algorithm, the Discrete Wavelet Transform (DWT), Principal Component Analysis (PCA), and an unsupervised K-means clustering for thresholding, feature extraction, to reduce features and for classification, respectively. Kernel-SVM machine learning algorithm is used for Classification. The comparative analysis of features extraction using DWT + KSVM, DWT + Otsu + KSVM, and DWT + Otsu + Kmeans + KSVM is done. The proposed model DWT + Otsu + Kmeans + KSVM gives promising results compared to the prior two approaches.KeywordsOtsuDiscrete wavelet transformK-means clusteringKernel-support vector machine
Sentiment Analysis (SA) is the current field of research in text mining field. SA is detecting opinions, sentiments, and subjectivity of text. It is the application of natural language processing techniques and text analytics to identify and extract subjective information from the frequently used sources such as web and microblogs. The main objective of sentiment analysis is to analyse reviews of products and services, and determine the scores of such sentiments. The major problem is that the reviews are mostly unstructured and thus, need classification or clustering to provide meaningful information for future use. This research work presents a survey of several machine learning techniques to enhance the classification accuracy in child YouTube data sentimental analysis. The hybrid Support vector machine model with ant colony optimization technique is employed for improving classification accuracy in child YouTube data sentiment analysis. The proposed hybrid classifier uses the results of the work compared with Naïve Bayes, SVM and Adaboosting + SVM classification techniques. The prediction for YouTube child input test sentences by each classifier is taken and the final output prediction is declared the one that has received harmful and secures decision. The proposed hybrid approach gives better accuracy in classification than the individual machine learning algorithms and also the existing proposed hybrid methods.
Dementia is a brain condition that impairs the cognitive abilities of an individual. Mild cognitive impairment is a mediator phase of healthy and dementia controls. The motivation of this study is to predict dementia using magnetic resonance imaging data, which is significant for the diagnosis of normal control and dementia patients. The proposed model leverages effective methods like Discrete Wavelet Transform, Bag of Features, and Support Vector Machine. The four wavelets haar, Daubechies, symlets, and coiflets are used for image compression. The results of the proposed data intelligence model are promising in terms of accuracy which is 92.32% which is better than the recently proposed models. Also, the proposed data intelligence model is compared with the models which may use curvelet transform, and shearlet transform and with the methods which have gone without using DWT transforms. The comparisons have also been made with the models that have used other prevalent techniques like Principal Component Analysis, Fisher Discriminant Ratio, and Gray Level Co-occurrence Matrix. The outcomes support the usage of each technique in the suggested data intelligence paradigm.
Medical anomaly identification using machine learning is a significant subject that has received a lot of attention. Artificial neural networks’ successor, deep learning, is a well-developed technology with strong computational capabilities. Its popularity has increased in recent years due to the availability of rapid data storage and hardware parallelism. Numerous, sizeable medical imaging datasets have recently been made available to the public, which has sparked interest in the field and increased the number of research studies and publications. The main goal of this study is to give a complete theoretical examination of prominent deep learning algorithms for detecting medical anomalies. The study further presents the architecture of current methodologies, compare and contrasts training algorithms, and gives a robust assessment of current methodologies. A thorough analysis of the state-of-the-art is provided, covering the benefits and limitations associated with using open-source data, and the specifications for clinically relevant systems. This study further identifies the gaps in the body of existing knowledge and suggests future research directions.
Dementia is a neurocognitive brain disease that emerged as a worldwide health challenge. Machine learning and deep learning have been effectively applied for the detection of dementia using magnetic resonance imaging. In this work, the performance of both machine learning and deep learning frameworks along with artificial neural networks are assessed for detecting dementia and normal subjects using MRI images. The first-order and second-order hand-crafted features are used as input for machine learning and artificial neural networks. And automatic feature extraction is used in the last framework with the pre-trained networks. The outcomes show that the framework using the deep neural networks performs better contrasted with the first two methodologies used in terms of various performance measures.
Each government is looking for to provide the best services to establish efficiency and quality of performance. This goal could be accomplished by improving the service performance of entire sectors in society. The government of Syria has realized the importance of moving in the direction of information technology. Therefore, E-governance initiatives were launched in Syria as a part of overall country information technology in 20 s century. Each government sector has since upgraded the performance by having its websites and e-services application. However, there are gaps and loose connections exist among the sectors, which has accordingly tarnished the image of Syrian E-governance. This has led to significant questions about the requirement of modification and enhancement of such service. Hence, the purpose of this research is to investigate and explore the factors that drive the E-governance implementation and affect government performance as well as the government-citizen relationship in Syria.KeywordsE-governanceGovernment of SyriaGovernment-Citizen
The adoption of microwave imaging as a tool for non-invasive monitoring of brain stroke has recently gained increasing attention. In this respect, the paper aims at providing a twofold contribution. First, we introduce a simple design tool to devise guidelines to properly set the working frequency as well as to choose the optimum matching medium needed to facilitate the penetration of the probing wave into the head. Second, we propose an imaging strategy based on a modified formulation of the linear sampling method, which allows a quasi real time monitoring of the disease's evolution. The accuracy of the design guidelines and performance of the imaging strategy are assessed through numerical examples dealing with 2D anthropomorphic phantoms.
In this paper, a novel approach based on the support vector machine (SVM) for dielectric target detection in through-wall scenario is proposed. Through-wall detection is converted to the establishment and use of a mapping between backscattered data and the dielectric parameter of the target. Then the propagation effects caused by walls, such as refraction and speed change, are included in the mapping that can be regressed after SVM training process. The training and testing data for the SVM is obtained by finite-difference time-domain (FDTD) simulation. Numerical experiments show that once the training phase is completed, this technique only needs computational time in an order of seconds to predict the parameters. Besides, experimental results show that good consistency between the actual parameters and estimated ones is achieved. Through-wall target tracking is also discussed and the results are acceptable.
In machining, the process of modeling and optimization are challenging tasks and need proper approaches to qualify the requirements in order to produce high quality of products with less cost estimation. There are a lot of modeling techniques that have been discovered by researches. In the recent years the trends were towards modeling of machining using computational approaches such as support vector machine (SVM), artificial neural network (ANN), genetic algorithm (GA), artificial bee colony (ACO) and particle swarm optimization (PSO). This paper reviews the application of SVM, classified as one of the popular trends in modeling techniques for both types of machining operations, conventional and modern machining. Generally, support vector machine is a powerful mathematical tool for data classification, regression and function estimation and also widely used for modeling machining operations. In SVM, there are several types of kernel function that used in SVM training parameters such as linear, polynomial, radial basis function (RBF), sigmoid and Gaussian kernel function. Review shows that RBF kernel function was widely applied in SVM as a kernel function in modeling machining performances.
Abstract—We have investigated the scattering of the Magnetic Resonance Imaging (MRI) radiofrequency (RF) field by implants for Deep Brain Stimulation (DBS) and the resultant heating of the tissue surrounding the DBS electrodes. The finite element method,has been used to perform full 3-D realistic simulations. The near field has been computed,for varying distances of the connecting portion of the lead from the air-tissue interface. Dissipated powers and induced temperature rise distributions have been obtained in the region surrounding the electrodes. It is shown that the near proximity of the air-tissue interface results in a reduction in the induced temperature rise.
Automatic classification of foreign fibers in cotton lint using machine vision is still a challenge due to various colors and shapes of the foreign fibers. This paper presents a novel classification method based on multi-class support vector machine (MSVM) which aims at accurate and fast classification of the foreign fibers. Firstly, live images were acquired by a machine vision system and then processed using image processing algorithms. Then the color features, shape features and texture features of each foreign fiber object were extracted and feature vectors were composed. Afterwards, three kinds of multi-class support vector machines were constructed, i.e., one-against-all decision-tree based MSVM, one-against-one voting based MSVM and one-against-one directed acyclic graph MSVM separately. At last, with the extracted feature vectors as input, the MSVMs were tested using leave-one-out cross validation. The results indicate that both the one-against-one voting based MSVM and the one-against-one directed acyclic graph MSVM can satisfy the accuracy requirement of the classification of foreign fibers, and the mean accuracy is 93.57% and 92.34% separately. The one-against-all decision-tree based MSVM only obtains mean accuracy of 79.25% which can not meet the accuracy requirement. In classification speed, one-against-one directed acyclic graph MSVM is the fastest and fitter for online classification.
Fractional Fourier domains form a continuum of domains making arbitrary angles with the time or frequency domains on the time–frequency plane. Signal representations in these domains are related to the fractional Fourier transform (FrFT). In this paper, a new proof on the shift-invariance of linear time–frequency distributions on fractional Fourier domains is given. We show that short-time Fourier transform (STFT) is the unique linear distribution satisfying magnitude-wise shift-invariance in the fractional Fourier domains. The magnitude-wise shift-invariance property in arbitrary fractional Fourier domains distinguishes STFT among all linear time–frequency distributions and simplifies the interpretation of the resultant distribution as shown by numerical examples.
In this paper, a differential wavelet-based operator defined on an interval is presented and used in evaluating the electromagnetic field described by Maxwell's curl equations, in time domain. The wavelet operator has been generated by using Daubechies wavelets with boundary functions. A spatial differential scheme has been performed and it has been applied in studying electromagnetic phenomena in a lossless medium. The proposed approach has been successfully tested on a bounded axial-symmetric cylindrical domain.
Present study examines biological effects of 2.45 GHz microwave radiation in Parkes strain mice. Forty-day-old mice were exposed to CW (continuous wave) microwave radiation (2 h/day for 30 days). Locomotor activity was recorded on running wheel for 12 days prior to microwave exposure (pre-exposure), 7 days during the first week of exposure (short-term exposure) and another 7-day spell during the last week of the 30-day exposure period (long-term exposure). Morris water maze test was performed from 17th to 22nd day of exposure. At the termination of the exposure, blood was processed for hematological parameters, brain for comet assay, epididymis for sperm count and motility and serum for SGOT (serum glutamate oxaloacetate transaminase) and SGPT (serum glutamate
An approach based on acoustics and its theoretical analogies to electromagnetism is used in the present research to study the detection of the acoustic wave energy radiated by the thermal random motion of material particles of the brain during activation or caused by pathology. Pressure and particle velocity are calculated in analytical mathematical forms for the case of human brain monitoring, which can be implemented by a prototype passive acoustic brain monitoring system (PABMOS). A sphere to model the human head and an internal point source in order to simulate potential pressure alterations due to intracranial abnormalities or local functional activations, are used in the theoretical representation of the present approach. Finally, numerical results for arbitrary positions of the internal source, concerning the particle velocity (pressure field distribution) at the surface of the head model which can implicitly be measured by the suitable piezoelectric sensors, are presented.
This paper presents a hybrid technique for the classification of the magnetic resonance images (MRI). The proposed hybrid technique consists of three stages, namely, feature extraction, dimensionality reduction, and classification. In the first stage, we have obtained the features related to MRI images using discrete wavelet transformation (DWT). In the second stage, the features of magnetic resonance images have been reduced, using principal component analysis (PCA), to the more essential features. In the classification stage, two classifiers have been developed. The first classifier based on feed forward back-propagation artificial neural network (FP-ANN) and the second classifier is based on k-nearest neighbor (k-NN). The classifiers have been used to classify subjects as normal or abnormal MRI human images. A classification with a success of 97% and 98% has been obtained by FP-ANN and k-NN, respectively. This result shows that the proposed technique is robust and effective compared with other recent work.
This paper describes a new method to classify sleep apnea syndrome (SAS) by using wavelet transforms and an artificial neural network (ANN). The network was trained and tested for different momentum coefficients. The abdominal respiration signals are separated into spectral components by using multi-resolution wavelet transforms. These spectral components are applied to the inputs of the artificial neural network. Then the neural network was configured to give three outputs to classify the SAS situation of the patient.The apnea can be broadly classified into three types: obstructive sleep apnea (OSA), central sleep apnea (CSA) and mixed sleep apnea (MSA). During OSA, the airway is blocked while respiratory efforts continue. During CSA the airway is open, however, there are no respiratory efforts. In this paper we aim to classify sleep apnea in one of three basic types: obstructive, central and mixed.
This paper proposes a hybrid crop classifier for polarimetric synthetic aperture radar (SAR) images. The feature sets consisted of span image, the H/A/α decomposition, and the gray-level co-occurrence matrix (GLCM) based texture features. Then, the features were reduced by principle component analysis (PCA). Finally, a two-hidden-layer forward neural network (NN) was constructed and trained by adaptive chaotic particle swarm optimization (ACPSO). K-fold cross validation was employed to enhance generation. The experimental results on Flevoland sites demonstrate the superiority of ACPSO to back-propagation (BP), adaptive BP (ABP), momentum BP (MBP), Particle Swarm Optimization (PSO), and Resilient back-propagation (RPROP) methods. Moreover, the computation time for each pixel is only 1.08 × 10⁻⁷ s.
Toe-walking is one of the most prevalent gait deviations and has been linked to many diseases. Three major ankle kinematic patterns have been identified in toe-walkers, but the relationships between the causes of toe-walking and these patterns remain unknown. This study aims to identify these relationships. Clearly, such knowledge would increase our understanding of this gait deviation, and could help clinicians plan treatment. The large quantity of data provided by gait analysis often makes interpretation a difficult task. Artificial intelligence techniques were used in this study to facilitate interpretation as well as to decrease subjective interpretation. Of the 716 limbs evaluated, 240 showed signs of toe-walking and met inclusion criteria. The ankle kinematic pattern of the evaluated limbs during gait was assigned to one of three toe-walking pattern groups to build the training data set. Toe-walker clinical measurements (range of movement, muscle spasticity and muscle strength) were coded in fuzzy modalities, and fuzzy decision trees were induced to create intelligible rules allowing toe-walkers to be assigned to one of the three groups. A stratified 10-fold cross validation situated the classification accuracy at 81%. Twelve rules depicting the causes of toe-walking were selected, discussed and characterized using kinematic, kinetic and EMG charts. This study proposes an original approach to linking the possible causes of toe-walking with gait patterns.
Abstract Learning is key to developing systems tailored to a broad range of data analysis and information extraction tasks. We outline the mathematical foundations of learning theory and describe a key algorithm of it.
In the history of research of the learning problem one can extract four periods that can be characterized by four bright events: (i) Constructing the first learning machines, (ii) constructing the fundamentals of the theory, (iii) constructing neural networks, (iv) constructing the alternatives to neural networks.
In this paper, a novel Wavelet-Galerkin Method (WGM) is presented to model the radio-wave propagation in tropospheric ducts. Galerkin method, with Daubechies scaling functions, is used to discretize the height operator. Later, a marching algorithm is developed using Crank-Nicolson (CN) method. A new"fictitious domain method"is also developed for parabolic wave equation to incorporate the impedance boundary conditions in WGM. In the end, results are compared with those from Advance Refractive Effects Prediction System (AREPS). Results show that the wavelet based methods are indeed feasible to model the radio wave propagation in troposphere as accurately as AREPS and proposed method can be a good alternative to other conventional methods.
During Magnetic Resonance Imaging (MRI), the presence of an implant such as a Deep Brain Stimulation (DBS) lead in a patient's body can pose a significant risk. This is due to the fact that the MR radiofrequency (RF) field can achieve a very high strength around the DBS electrodes. Thus the specific absorption rate (SAR), which is proportional to the square of the magnitude of the RF electric field, can have a very high concentration in the near-field region of the electrodes. The resulting tissue heating can reach dangerous levels. The degree of heating depends on the level of SAR concentration. The effects can be severe, leading to tissue ablation and brain damage, and significant safety concerns arise whenever a patient with an implanted DBS lead is exposed to MR scanning. In this paper, SAR, electric field, and temperature rise distributions have been found around actual DBS electrodes. The magnitude and spatial distribution of the induced temperature rises are found to be a function of the length and structure of the lead device, tissue properties and the MR stimulation paramete
During the past decades there has been a tremendous increase throughout the scientific community for developing methods of understanding human brain functionality, as diagnosis and treatment of diseases and malfunctions, could be effectively developed through understanding of how the brain works. In parallel, research effort is driven on minimizing drawbacks of existing imaging techniques including potential risks from radiation and invasive attributes of the imaging methodologies. Towards that direction a new near field radiometry imaging system has been theoretically studied, developed and experimentally tested and all of the aforementioned research phases are herein presented. The system operation principle is based on the fact that human tissues emit chaotic thermal type radiation at temperatures above the absolute zero. Using a phase shifted antenna array system, spatial resolution, detection depth and sensitivity are increased. Combining previous research results, as well as new findings, the capabilities of the constructed system, as well as the possibility of using it as a complementary method for brain imaging are discussed in this paper.
Automated and accurate classification of magnetic resonance (MR) brain images is an integral component of the analysis and interpretation of neuroimaging. Many different and innovative methods have been proposed to improve upon this technology. In this study, we presented a forward neural network (FNN) based method to classify a given MR brain image as normal or abnormal. This method first employs a wavelet transform to extract features from images, and then applies the technique of principle component analysis (PCA) to reduce the dimensions of features. The reduced features are sent to an FNN, and these parameters are optimized via adaptive chaotic particle swarm optimization (ACPSO). K-fold stratified cross validation was used to enhance generalization. We applied the proposed method on 160 images (20 normal, 140 abnormal), and found that the classification accuracy is as high as 98.75% while the computation time per image is only 0.0452s.
Automated and accurate classification of magnetic resonance (MR) brain images is a hot topic in the field of neuroimaging. Recently many different and innovative methods have been proposed to improve upon this technology. In this study, we presented a hybrid method based on forward neural network (FNN) to classify an MR brain image as normal or abnormal. The method first employed a discrete wavelet transform to extract features from images, and then applied the technique of principle component analysis (PCA) to reduce the size of the features. The reduced features were sent to an FNN, of which the parameters were optimized via an improved artificial bee colony (ABC) algorithm based on both fitness scaling and chaotic theory. We referred to the improved algorithm as scaled chaotic artificial bee colony (SCABC). Moreover, the K-fold stratified cross validation was employed to avoid overfitting. In the experiment, we applied the proposed method on the data set of T2-weighted MRI images consisting of 66 brain images (18 normal and 48 abnormal). The proposed SCABC was compared with traditional training methods such as BP, momentum BP, genetic algorithm, elite genetic algorithm with migration, simulated annealing, and ABC. Each algorithm was run 20 times to reduce randomness. The results show that our SCABC can obtain the least mean MSE and 100% classification accuracy.
In this paper we explain how to characterize the best approximation to any xx in a Hilbert space XX from the set C∩{x∈X:gi(x)≤0,i=1,2,…,m}C∩{x∈X:gi(x)≤0,i=1,2,…,m} in the face of data uncertainty in the convex constraints, gi(x)≤0,i=1,2,…,mgi(x)≤0,i=1,2,…,m, where CC is a closed convex subset of XX. Following the robust optimization approach, we establish Lagrange multiplier characterizations of the robust constrained best approximation that is immunized against data uncertainty. This is done by characterizing the best approximation to any xx from the robust counterpart of the constraints where the constraints are satisfied for all possible uncertainties within the prescribed uncertainty sets. Unlike the traditional Lagrange multiplier characterizations without data uncertainty, for constrained best approximation problems in the face uncertainty, we show that the strong conical hull intersection property (strong CHIP) alone is not sufficient to guarantee the Lagrange multiplier characterizations. We present conditions which guarantee that the strong CHIP is necessary and sufficient for the multiplier characterization. We also establish that the strong CHIP is automatically satisfied for the cases of polyhedral constraints with polytope uncertainty, and linear constraints with interval uncertainty. As an application, we show how robust solutions of shape preserving interpolation problems under ellipsoidal and box uncertainty cases can be obtained in terms of Lagrange multipliers under strict robust feasibility conditions.
Feed forward neural Network (FNN) has been widely applied to many fields because of its ability to closely approximate unknown function to any degree of desired accuracy. Back Propagation (BP) is the most general learning algorithms, but is subject to local optimal convergence and poor performance even on simple problems when forecasting out of samples. Thus, we proposed an improved Bacterial Chemotaxis Optimization (BCO) approach as a possible alternative to the problematic BP algorithm, along with a novel adaptive search strategy to improve the efficiency of the traditional BCO. Taking the classical XOR problem and sinc function approximation as examples, comparisons were implemented. The results demonstrate that our algorithm is obviously superior in convergence rate and precision compared with other training algorithms, such as Genetic Algorithm (GA) and Taboo Search (TS).
Automated animal behaviour monitoring systems have become increasingly appealing for research and animal production management purposes. However, many existing systems are suited to measure only one or two behaviour patterns or activity states at a time. We aimed to develop and pilot a method for automatically measuring and recognising several behavioural patterns of dairy cows using a three-dimensional accelerometer and a multi-class support vector machine (SVM). SVM classification models were constructed based on nine features. The models were trained using observations made of the behaviour of 30 cows fitted with a neck collar bearing an accelerometer that recorded horizontal, vertical and lateral acceleration. Measured behaviour patterns included standing, lying, ruminating, feeding, normal and lame walking, lying down, and standing up. Accuracy, sensitivity, precision, and kappa measures were used to evaluate the model performance. The SVM classification models achieved a reasonable recognition of standing (80% sensitivity, 65% precision), lying (80%, 83%), ruminating (75%, 86%), feeding (75%, 81%), walking normally (79%, 79%), and lame walking (65%, 66%). The results were poor for lying down (0%, 0%) and standing up (71%, 29%). The overall performance of the multi-class model was 78% precision with a kappa value of 0.69. Each of the behaviour categories had one or two other behaviour patterns that became confused with them the most. The problematic behaviours were expectedly those that resemble each other in terms of movement. Possible solutions for the problems in classification are presented. In conclusion, accelerometers can be used to easily recognise various behaviour patterns in dairy cows. Support vector machines proved useful in classification of measured behaviour patterns. However, further work is needed to refine the features used in the classification models in order to gain the best possible classification performance. Also the quality of acceleration data needs to be considered to improve the results.
This paper proposes a hybrid classifier for polarimetric SAR images. The feature sets consist of span image, the H/A/α decomposition, and the gray-level co-occurrence matrix (GLCM) based texture features. Then, the features are reduced by principle component analysis (PCA). A 3-layer neural network (NN) is constructed, trained by resilient back-propagation (RPROP) method to fasten the training and early stop (ES) method to prevent the overfitting. The results of San Francisco and Flevoland sites compared to Wishart Maximum Likelihood and wavelet-based method demonstrate the validness of our method in terms of confusion matrix and overall accuracy. In addition, NNs with and without PCA are compared. Results show the NN with PCA is more accurate and faster.
Least Square Support Vector Machines (LS-SVM), a new machine-learning tool has been employed for developing data driven models of non-linear processes. The method is firmly rooted in the statistical learning theory and transforms the input data to a higher dimensional feature space where the use of appropriate kernel functions avoid computational difficulty. Further, a pattern search algorithm, which explores multiple directions and utilizes coordinate search with fixed step size, is employed for selecting optimal LS-SVM model that produces a minimum possible prediction error. To show the efficacy and efficiency of the fully automated pattern search assisted LS-SVM methodology, we have tested it on several benchmark examples. The study suggests that proposed paradigm can be a useful and viable tool in building data driven models of non-linear processes.
Setting of the learning problem consistency of learning processes bounds on the rate of convergence of learning processes controlling the generalization ability of learning processes constructing learning algorithms what is important in learning theory?.
An approach for the microwave nonlinear device modeling technique based on a combination of the conventional equivalent circuit model and support vector machine (SVM) regression is presented in this paper. The intrinsic nonlinear circuit elements are represented by Taylor series expansions, coefficients of which are predicted by its support vector regression (SVR) model. Example of a SiC MESFET nonlinear model is demonstrated, and good results is achieved.
This manuscript aims at illustrating significant refinements concerning the use of wavelets, when these latter are used in the guise of continuous wavelet transforms (CWT) for identifying damage on transversally vibrating structural components (e.g. beams, plates and shells). The refinements regard the presentation of wavelet-algorithms which are aimed at significantly reducing those border distortions normally arising during a wavelet-damage detection procedure. The main advantage of the algorithms is that they are self-contained, namely: (i) the wavelet transforms do not undergo any own variation and their application follows the convolution laws established in the past; (ii) it is not necessary to design a specific boundary wavelet; (iii) no significant analytical treatment neither of the wavelets nor of the signal is required and, finally, (iv) the algorithms can be adapted to different boundary conditions and different physical situations. Besides all the specified advantages, the wavelet-damage detection procedure is still carried out by excluding historical data. The effectiveness of the algorithms is shown through numerical and experimental examples. These latter are illustrated along with reduced outliers of experimental estimation through a related consistent statistical procedure.
Deep brain stimulation (DBS) is a well-established treatment for Parkinson's disease, essential tremor and dystonia. It has also been successfully applied to treat various other neurological and psychiatric conditions including depression and obsessive-compulsive disorder. Numerous computational models, mostly based on the Finite Element Method (FEM) approach have been suggested to investigate the biophysical mechanisms of electromagnetic wave-tissue interaction during DBS. These models, although emphasizing the importance of various electrical and geometrical parameters, mostly have used simplified geometries over a tightly restricted tissue volume in the case of monopolar stimulation. In the present work we show that topological arrangements and geometrical properties of the model have a significant effect on the distribution of voltages in the concerned tissues. The results support reconsidering the current approach for modeling monopolar DBS which uses a restricted cubic area extended a few centimeters around the active electrode to predict the volume of activated tissue. We propose a new technique called multi-resolution FEM modeling, which may improve the accuracy of the prediction of volume of activated tissue and yet be computationally tractable on personal computers.
A new method of reducing the computational load in decision functions provided by a support vector classification machine is studied. The method exploits the geometrical relations when the kernels used are based on distances to obtain bounds of the remaining decision function and avoids to continue calculating kernel operations when there is no chance to change the decision. The method proposed achieves savings in operations of 25–90% whilst keeping the same accuracy. Although the method is explained for support vector machines, it can be applied to any kernel binary classifier that provides a similar evaluation function.
A main theme of this report is the relationship of approximation to learning and the primary role of sampling (inductive inference). We try to emphasize relations of the theory of learning to the main stream of mathematics.
Hitherto communication theory was based on two alternative methods of signal analysis. One is the description of the signal as a function of time; the other is Fourier analysis. Both are idealizations, as the first method operates with sharply defined instants of time, the second with infinite wave-trains of rigorously defined frequencies. But our everyday experiences¿especially our auditory sensations¿insist on a description in terms of both time and frequency. In the present paper this point of view is developed in quantitative language. Signals are represented in two dimensions, with time and frequency as co-ordinates. Such two-dimensional representations can be called ¿information diagrams,¿ as areas in them are proportional to the number of independent data which they can convey. This is a consequence of the fact that the frequency of a signal which is not of infinite duration can be defined only with a certain inaccuracy, which is inversely proportional to the duration, and vice versa. This ¿uncertainty relation¿ suggests a new method of description, intermediate between the two extremes of time analysis and spectral analysis. There are certain ¿elementary signals¿ which occupy the smallest possible area in the information diagram. They are harmonic oscillations modulated by a ¿probability pulse.¿ Each elementary signal can be considered as conveying exactly one datum, or one ¿quantum of information.¿ Any signal can be expanded in terms of these by a process which includes time analysis and Fourier analysis as extreme cases. These new methods of analysis, which involve some of the mathematical apparatus of quantum theory, are illustrated by application to some problems of transmission theory, such as direct generation of single sidebands, signals transmitted in minimum time through limited frequency channels, frequency modulation and time-division multiplex telephony.
Otoliths have traditionally been used to estimate fish age. However, many factors influence changes in otolith shape, so manual classification remains a complicated task. Very recently, statistical learning techniques have been proposed for automating such a process. We propose performing automatic fish age classification using otolith images (in cases in which growth rings are not properly displayed or are unavailable), morphological and statistical feature-extraction methods and multi-class support vector machines. The results of our experiments, in which we classified cod ages from otolith images, demonstrate the effectiveness of the approach.
In this paper, we propose a novel method using wavelets as input to neural network self-organizing maps and support vector machine for classification of magnetic resonance (MR) images of the human brain. The proposed method classifies MR brain images as either normal or abnormal. We have tested the proposed approach using a dataset of 52 MR brain images. Good classification percentage of more than 94% was achieved using the neural network self-organizing maps (SOM) and 98% from support vector machine. We observed that the classification rate is high for a support vector machine classifier compared to self-organizing map-based approach.
Support Vector Machines (SVMs) have achieved very good performance on different learning problems. However, the success of SVMs depends on the adequate choice of the values of a number of parameters (e.g., the kernel and regularization parameters). In the current work, we propose the combination of meta-learning and search algorithms to deal with the problem of SVM parameter selection. In this combination, given a new problem to be solved, meta-learning is employed to recommend SVM parameter values based on parameter configurations that have been successfully adopted in previous similar problems. The parameter values returned by meta-learning are then used as initial search points by a search technique, which will further explore the parameter space. In this proposal, we envisioned that the initial solutions provided by meta-learning are located in good regions of the search space (i.e. they are closer to optimum solutions). Hence, the search algorithm would need to evaluate a lower number of candidate solutions when looking for an adequate solution. In this work, we investigate the combination of meta-learning with two search algorithms: Particle Swarm Optimization and Tabu Search. The implemented hybrid algorithms were used to select the values of two SVM parameters in the regression domain. These combinations were compared with the use of the search algorithms without meta-learning. The experimental results on a set of 40 regression problems showed that, on average, the proposed hybrid methods obtained lower error rates when compared to their components applied in isolation.
Magnetic resonance imaging (MRI) segmentation has been implemented by many clustering techniques, such as k-means, fuzzy c-means (FCM), learning-vector quantization (LVQ) and fuzzy algorithms for LVQ (FALVQ). Although these algorithms have been successful in applying MRI segmentation, obtaining the right number of clusters and adapting to different cluster characteristics are still not satisfactorily addressed. This report proposes an optimization technique, a hierarchical genetic algorithm with a fuzzy learning-vector quantization network (HGALVQ), to segment multi-spectral human-brain MRI. Evaluation of this approach is based on a real case with human-brain MRI of an individual suffering from meningioma. The HGALVQ is verified by the comparison with other popular clustering algorithms such as k-means, FCM, FALVQ, LVQ, and simulated annealing. Experimental results show that HGALVQ not only returns an appropriate number of clusters and also outperforms other methods in specificity.
In nonparametric classification and regression problems, regularized kernel
methods, in particular support vector machines, attract much attention in
theoretical and in applied statistics. In an abstract sense, regularized kernel
methods (simply called SVMs here) can be seen as regularized M-estimators for a
parameter in a (typically infinite dimensional) reproducing kernel Hilbert
space. For smooth loss functions, it is shown that the difference between the
estimator, i.e.\ the empirical SVM, and the theoretical SVM is asymptotically
normal with rate . That is, the standardized difference converges
weakly to a Gaussian process in the reproducing kernel Hilbert space. As common
in real applications, the choice of the regularization parameter may depend on
the data. The proof is done by an application of the functional delta-method
and by showing that the SVM-functional is suitably Hadamard-differentiable.
Data splitting is an important consideration during artificial neural network (ANN) development where hold-out cross-validation is commonly employed to ensure generalization. Even for a moderate sample size, the sampling methodology used for data splitting can have a significant effect on the quality of the subsets used for training, testing and validating an ANN. Poor data splitting can result in inaccurate and highly variable model performance; however, the choice of sampling methodology is rarely given due consideration by ANN modellers. Increased confidence in the sampling is of paramount importance, since the hold-out sampling is generally performed only once during ANN development. This paper considers the variability in the quality of subsets that are obtained using different data splitting approaches. A novel approach to stratified sampling, based on Neyman sampling of the self-organizing map (SOM), is developed, with several guidelines identified for setting the SOM size and sample allocation in order to minimize the bias and variance in the datasets. Using an example ANN function approximation task, the SOM-based approach is evaluated in comparison to random sampling, DUPLEX, systematic stratified sampling, and trial-and-error sampling to minimize the statistical differences between data sets. Of these approaches, DUPLEX is found to provide benchmark performance with good model performance, with no variability. The results show that the SOM-based approach also reliably generates high-quality samples and can therefore be used with greater confidence than other approaches, especially in the case of non-uniform datasets, with the benefit of scalability to perform data splitting on large datasets.
Deep brain stimulation of the subthalamic nucleus (DBS-STN) is an approved treatment for advanced Parkinson disease (PD) patients; however, there is a need to further evaluate its effect on gait. This study compares logistic regression (LR), probabilistic neural network (PNN) and support vector machine (SVM) classifiers for discriminating between normal and PD subjects in assessing the effects of DBS-STN on ground reaction force (GRF) with and without medication. Gait analysis of 45 subjects (30 normal and 15 PD subjects who underwent bilateral DBS-STN) was performed. PD subjects were assessed under four test conditions: without treatment (mof-sof), with stimulation alone (mof-son), with medication alone (mon-sof), and with medication and stimulation (mon-son). Principal component (PC) analysis was applied to the three components of GRF separately, where six PC scores from vertical, one from anterior-posterior and one from medial-lateral were chosen by the broken stick test. Stepwise LR analysis employed the first two and fifth vertical PC scores as input variables. Using the bootstrap approach to compare model performances for classifying GRF patterns from normal and untreated PD subjects, the first three and the fifth vertical PCs were attained as SVM input variables, while the same ones plus the first anterior-posterior were selected as PNN input variables. PNN performed better than LR and SVM according to area under the receiver operating characteristic curve and the negative likelihood ratio. When evaluating treatment effects, the classifiers indicated that DBS-STN alone was more effective than medication alone, but the greatest improvements occurred with both treatments together.
A new model for neuro-fuzzy (NF) classification systems is proposed. The motivation is to utilize the feature-wise degree of belonging of patterns to all classes that are obtained through a fuzzification process. A fuzzification process generates a membership matrix having total number of elements equal to the product of the number of features and classes present in the data set. These matrix elements are the input to neural networks. The effectiveness of the proposed model is established with four benchmark data sets (completely labeled) and two remote sensing images (partially labeled). Different performance measures such as misclassification, classification accuracy and kappa index of agreement for completely labeled data sets, and beta index of homogeneity and Davies-Bouldin (DB) index of compactness for remotely sensed images are used for quantitative analysis of results. All these measures supported the superiority of the proposed NF classification model. The proposed model learns well even with a lower percentage of training data that makes the system fast.
A novel, fully automatic, adaptive, robust procedure for brain tissue classification from 3D magnetic resonance head images (MRI) is described in this paper. The procedure is adaptive in that it customizes a training set, by using a 'pruning' strategy, such that the classification is robust against anatomical variability and pathology. Starting from a set of samples generated from prior tissue probability maps (a 'model') in a standard, brain-based coordinate system ('stereotaxic space'), the method first reduces the fraction of incorrectly labeled samples in this set by using a minimum spanning tree graph-theoretic approach. Then, the corrected set of samples is used by a supervised kNN classifier for classifying the entire 3D image. The classification procedure is robust against variability in the image quality through a non-parametric implementation: no assumptions are made about the tissue intensity distributions. The performance of this brain tissue classification procedure is demonstrated through quantitative and qualitative validation experiments on both simulated MRI data (10 subjects) and real MRI data (43 subjects). A significant improvement in output quality was observed on subjects who exhibit morphological deviations from the model due to aging and pathology.
- C M Bishop
Bishop, C. M., Pattern Recognition and Machine Learning
(Information Science and Statistics), Springer-Verlag New York,
Inc., 2006.
Hybrid intelligent techniques for MRI brain images classification
- E.-S A El-Dahshan
- T Hosny
- A.-B M Salem
El-Dahshan, E.-S. A., T. Hosny, and A.-B. M. Salem, "Hybrid intelligent techniques for MRI
brain images classification," Digital Signal Processing, Vol. 20, No. 2, pp. 433-441, 2010.
2.45 GHz (Cw) Microwave Irradiation Alters Circadian Organization, Spatial Memory, Dna Structure in the Brain Cells and Blood Cell Counts of Male Mice, Mus Musculus
- C M Chaturvedi
Chaturvedi, C. M., et al., "2.45 GHz (Cw) Microwave Irradiation Alters Circadian
Organization, Spatial Memory, Dna Structure in the Brain Cells and Blood Cell Counts of
Male Mice, Mus Musculus," Progress In Electromagnetics Research B, Vol. 29, No. pp. 23-42,
2011.
- Chim
Chim. Acta Volume 642 (2009) 59-68]," Analytica Chimica Acta, Vol. 658, No. 1, pp. 106-106,
2010.
- C M Bishop
Bishop, C. M., Pattern Recognition and Machine Learning (Information Science and
Statistics): Springer-Verlag New York, Inc., 2006.