No full-text available
To read the full-text of this research,
you can request a copy directly from the author.
Multiscale adaptive and attention-dilated convolutional neural network for efficient leukemia detection model with multiscale trans-res-Unet3+-based segmentation network
Abstract
A precise early-stage leukemia diagnosis is essential for treating patients and saving their lives. The least expensive way for the initial diagnosis of leukemia in patients is the microscopy imaging analysis. However, this work is subjective and time-consuming. Hence, this paper creates an efficient leukemia detection model using deep learning approaches. Initially, the standard leukemia datasets are used to collect the images. The gathered images are given for the region segmentation to the Multiscale Trans-Res-Unet3+ (MTResUnet3+) Network. The segmented regions from the MTResUnet3+ are now considered for the feature extraction phase from which the most relevant attributes are mined. Here, the features like color, shape, and texture are extracted separately for performing efficient detection. Further, the features being extracted are considered for the feature selection phase, where the Election-Based Chameleon Swarm Algorithm (E-CSA) is utilized to optimally select the most appropriate features with the aim of enhancing the performance rate of the developed model. The optimally selected features are given to the next stage for detecting the presence of leukemia. Here, the Multiscale Adaptive and Attention-based Dilated Convolutional Neural Network (MAA-DCNN) is made for detecting leukemia, in which the optimization of the parameter is done with the help of hybrid E-CSA in order to elevate the detection accuracy of leukemia. The simulation analysis is performed to analyze the performance rate of the recommended leukemia detection model by contrasting it with the conventional leukemia detection models and existing algorithms using various performance metrics for validation.
... To overcome the inherent complexities of leukemia detection, researchers have proposed various deep learning architectures [8,9,10], including ensemble methods incorporating CNNs with recurrent architectures [11], stain deconvolutional CNNs with auxiliary classifiers [12], depthwise convolutions with varying dilation rates [13] and architectures specifically designed to handle class imbalance [14]. Beyond the deep learning-based architectures, Khandekar et al. [15] highlights the potential of YOLOv4 with feature enhancement techniques for leukemia detection. ...
... However, these methods encounter significant limitations. First, some existing techniques [8,15] attempt feature extraction directly from raw blood smears, introducing contamination from irrelevant elements, compromising diagnostic value for Acute Lymphoblastic Leukemia (ALL). Leukocyte segmentation isolates features derived solely from the cells of interest, leading to greater accuracy. ...
... Medical image segmentation is essential for the analysis and diagnosis of diseases, as it allows for the precise identification of Regions of Interest (ROIs) [23,24] within samples. Gokulkannan et al. [8] employed a ResNet-based UNet to segment leukocytes, isolating regions of interest for feature detection. The obtained features were optimized using the Election-Based Chameleon Swarm Algorithm (E-CSA). ...
Background:
Swift and accurate blood smear analyses are crucial for diagnosing leukemia and other hematological malignancies. However, manual leukocyte count and morphological evaluation remain time-consuming and prone to errors. Additionally, conventional image processing methods struggle to differentiate cells due to visual similarities between malignant and benign cell morphology.
Method:
In response to above challenges, we propose Coupled Transformer Convolutional Network (CoTCoNet) framework for leukemia classification. CoTCoNet integrates dual-feature extraction to capture long-range global features and fine-grained spatial patterns, facilitating the identification of complex hematological characteristics. Additionally, the framework employs a graph-based module to uncover hidden, biologically relevant features of leukocyte cells, along with a Population-based Meta Heuristic Algorithm for feature selection and optimization. Furthermore, we introduce a novel combination of leukocyte segmentation and synthesis, which isolates relevant regions while augmenting the training dataset with realistic leukocyte samples. This strategy isolates relevant regions while augmenting the training data with realistic leukocyte samples, enhancing feature extraction, and addressing data scarcity without compromising data integrity.
Results:
We evaluated CoTCoNet on a dataset of 16,982 annotated cells, achieving an accuracy of 0.9894 and an F1-Score of 0.9893. We tested CoTCoNet on four diverse, publicly available datasets (including those above) to assess generalizability. Results demonstrate a significant performance improvement over existing state-of-the-art approaches.
Conclusions:
CoTCoNet represents a significant advancement in leukemia classification, offering enhanced accuracy and efficiency compared to traditional methods. By incorporating explainable visualizations that closely align with cell annotations, the framework provides deeper insights into its decision-making process, further solidifying its potential in clinical settings.
... However, this method has the drawback of its inability to classify lymphoblast into various sub-types. This weakness was addressed in Pui et al. (2004), Döhner et al. (2015), Karim et al. (2021), Gupta et al. (2022), Mustaqim et al. (2023), Gokulkannan et al. (2024) and the modified methods are capable of classification into different subcategories of Leukaemia such as Acute Lymphocytic Leukaemia (ALL) and Acute Myeloid Leukaemia (AML). A study which was presented by Madhloom et al. (2012) mainly focused on detecting ALL where feature extraction is achieved using shape and texture features, normalization, and Fishers Discrimination ratio followed by the color and morphological reconstruction as well as an exhaustive search, and finally, KNN (Guo et al. 2003) based on Euclidean distances is used for classification. ...
About 300,000 leukemia cases are diagnosed every year, with the total number of active cases rising to 2.3 million in 2015. Although the number of adults diagnosed with leukemia is pretty high, this is the most common type of cancer found in children in developed countries. Its ability to recur and expensive diagnostic process make patients unable to undergo the diagnosis on a timely basis and consequently can prove fatal for many. The proposed novel model PyraNet aims to tackle the requirement of high-precision machinery and human expertise, as there might not be enough resources for the latter. Proportionate fine-tuning and construction make the model to accurately and precisely detect the presence of leukemic blast cells and classify them into their respective class types. Also, this novel architecture proposed here is a step towards solving the problem of unbalanced classes that often arises when the quantitative distribution of data within different classes is highly biased. As an initiative to tackle it, we have used multi-model-layer training. The analysis of experimental results shows that the proposed model is capable of correctly predicting a higher number of classes with better accuracy.
... Efficient segmentation methods markedly improve the diagnosis of acute lymphoblastic leukemia (ALL) by precisely delineating areas of interest in medical imaging, including bone marrow biopsies and blood smears. This procedure enhances classification precision by diminishing background noise, accentuating essential cellular characteristics, and standardizing diversity among patient presentations [79]. Segmentation improves feature extraction and enables multimodal analysis by supplying cleaner, more pertinent data, enhancing the integration of images, genomic, and clinical information. ...
Automatic detection of Leukemia or blood cancer is one of the most challenging tasks that need to be addressed in the healthcare system. Analysis of white blood cells (WBCs) in the blood or bone marrow microscopic slide images play a crucial part in early identification to facilitate medical experts. For Acute Lymphocytic Leukemia (ALL), the most preferred part of the blood or marrow is to be analyzed by the experts before it spreads in the whole body and the condition becomes worse. The researchers have done a lot of work in this field, to demonstrate a comprehensive analysis few literature reviews have been published focusing on various artificial intelligence-based techniques like machine and deep learning detection of ALL. The systematic review has been done in this article under the PRISMA guidelines which presents the most recent advancements in this field. Different image segmentation techniques were broadly studied and categorized from various online databases like Google Scholar, Science Direct, and PubMed as image processing-based, traditional machine and deep learning-based, and advanced deep learning-based models were presented. Convolutional Neural Networks (CNN) based on traditional models and then the recent advancements in CNN used for the classification of ALL into its subtypes. A critical analysis of the existing methods is provided to offer clarity on the current state of the field. Finally, the paper concludes with insights and suggestions for future research, aiming to guide new researchers in the development of advanced automated systems for detecting life-threatening diseases.
... Efficient segmentation methods markedly improve the diagnosis of acute lymphoblastic leukemia (ALL) by precisely delineating areas of interest in medical imaging, including bone marrow biopsies and blood smears. This procedure enhances classification precision by diminishing background noise, accentuating essential cellular characteristics, and standardizing diversity among patient presentations [79]. Segmentation improves feature extraction and enables multimodal analysis by supplying cleaner, more pertinent data, enhancing the integration of images, genomic, and clinical information. ...
Automatic detection of Leukemia or blood cancer is one of the most challenging tasks that need to be addressed in the healthcare system. Analysis of white blood cells (WBCs) in the blood or bone marrow microscopic slide images play a crucial part in early identification to facilitate medical experts. For Acute Lymphocytic Leukemia (ALL), the most preferred part of the blood or marrow is to be analyzed by the experts before it spreads in the whole body and the condition becomes worse. The researchers have done a lot of work in this field, to demonstrate a comprehensive analysis few literature reviews have been published focusing on various artificial intelligence-based techniques like machine and deep learning detection of ALL. The systematic review has been done in this article under the PRISMA guidelines which presents the most recent advancements in this field. Different image segmentation techniques were broadly studied and categorized from various online databases like Google Scholar, Science Direct, and PubMed as image processing-based, traditional machine and deep learning-based, and advanced deep learning-based models were presented. Convolutional Neural Networks (CNN) based on traditional models and then the recent advancements in CNN used for the classification of ALL into its subtypes. A critical analysis of the existing methods is provided to offer clarity on the current state of the field. Finally, the paper concludes with insights and suggestions for future research, aiming to guide new researchers in the development of advanced automated systems for detecting life-threatening diseases.
... Wang et al. [21] proposed a method that combines Fourier ptychographic microscopy with YOLO, enabling the acquisition of high-resolution, wide-field blood cell images in a single shot, thereby improving detection accuracy. K. Gokulkannan et al. [22] designed a multi-scale adaptive and attention-based DCNN (MAA-DCNN) approach and developed a new multi-scale Trans-Res-Unet3+ (MTResUnet3+) model using this method, enhancing the detection performance for pathological blood cells. ...
As deep learning technology has progressed, automated medical image analysis is becoming ever more crucial in clinical diagnosis. However, due to the diversity and complexity of blood cell images, traditional models still exhibit deficiencies in blood cell detection. To address blood cell detection, we developed the TW-YOLO approach, leveraging multi-scale feature fusion techniques. Firstly, traditional CNN (Convolutional Neural Network) convolution has poor recognition capabilities for certain blood cell features, so the RFAConv (Receptive Field Attention Convolution) module was incorporated into the backbone of the model to enhance its capacity to extract geometric characteristics from blood cells. At the same time, utilizing the feature pyramid architecture of YOLO (You Only Look Once), we enhanced the fusion of features at different scales by incorporating the CBAM (Convolutional Block Attention Module) in the detection head and the EMA (Efficient Multi-Scale Attention) module in the neck, thereby improving the recognition ability of blood cells. Additionally, to meet the specific needs of blood cell detection, we designed the PGI-Ghost (Programmable Gradient Information-Ghost) strategy to finely describe the gradient flow throughout the process of extracting features, further improving the model’s effectiveness. Experiments on blood cell detection datasets such as BloodCell-Detection-Dataset (BCD) reveal that TW-YOLO outperforms other models by 2%, demonstrating excellent performance in the task of blood cell detection. In addition to advancing blood cell image analysis research, this work offers strong technical support for future automated medical diagnostics.
... This work addresses the use of Self-Supervised Learning (SSL) in combination with multihead attention to categorize the subtype classification from blood smears. Gokulkannan et al. 44 presented a highly effective model for detecting leukaemia. The model utilizes the Multiscale Trans-Res-Unet3 + Network to segment the region of interest. ...
The bone marrow overproduces immature cells in the malignancy known as Acute Lymphoblastic Leukemia (ALL). In the United States, about 6500 occurrences of ALL are diagnosed each year in both children and adults, comprising nearly 25% of pediatric cancer cases. Recently, many computer-assisted diagnosis (CAD) systems have been proposed to aid hematologists in reducing workload, providing correct results, and managing enormous volumes of data. Traditional CAD systems rely on hematologists’ expertise, specialized features, and subject knowledge. Utilizing early detection of ALL can aid radiologists and doctors in making medical decisions. In this study, Deep Dilated Residual Convolutional Neural Network (DDRNet) is presented for the classification of blood cell images, focusing on eosinophils, lymphocytes, monocytes, and neutrophils. To tackle challenges like vanishing gradients and enhance feature extraction, the model incorporates Deep Residual Dilated Blocks (DRDB) for faster convergence. Conventional residual blocks are strategically placed between layers to preserve original information and extract general feature maps. Global and Local Feature Enhancement Blocks (GLFEB) balance weak contributions from shallow layers for improved feature normalization. The global feature from the initial convolution layer, when combined with GLFEB-processed features, reinforces classification representations. The Tanh function introduces non-linearity. A Channel and Spatial Attention Block (CSAB) is integrated into the neural network to emphasize or minimize specific feature channels, while fully connected layers transform the data. The use of a sigmoid activation function concentrates on relevant features for multiclass lymphoblastic leukemia classification The model was analyzed with Kaggle dataset (16,249 images) categorized into four classes, with a training and testing ratio of 80:20. Experimental results showed that DRDB, GLFEB and CSAB blocks’ feature discrimination ability boosted the DDRNet model F1 score to 0.96 with minimal computational complexity and optimum classification accuracy of 99.86% and 91.98% for training and testing data. The DDRNet model stands out from existing methods due to its high testing accuracy of 91.98%, F1 score of 0.96, minimal computational complexity, and enhanced feature discrimination ability. The strategic combination of these blocks (DRDB, GLFEB, and CSAB) are designed to address specific challenges in the classification process, leading to improved discrimination of features crucial for accurate multi-class blood cell image identification. Their effective integration within the model contributes to the superior performance of DDRNet.
Acute leukemia is characterized by the swift proliferation of immature white blood cells (WBC) in the blood and bone marrow. It is categorized into acute lymphoblastic leukemia (ALL) and acute myeloid leukemia (AML), depending on whether the cell-line origin is lymphoid or myeloid, respectively. Deep learning (DL) and artificial intelligence (AI) are revolutionizing medical sciences by assisting clinicians with rapid illness identification, reducing workload, and enhancing diagnostic accuracy. This paper proposes a DL-based novel BSNEU-net framework to detect acute leukemia. It comprises 4 Union Blocks (UB) and incorporates block feature map distortion (BFMD) with switchable normalization (SN) in each UB. The UB employs union convolution to extract more discriminant features. The BFMD is adapted to acquire more generalized patterns to minimize overfitting, whereas SN layers are appended to improve the model’s convergence and generalization capabilities. The uniform utilization of batch normalization across convolution layers is sensitive to the mini-batch dimension changes, which is effectively remedied by incorporating an SN layer. Here, a new dataset comprising 2400 blood smear images of ALL, AML, and healthy cases is proposed, as DL methodologies necessitate a sizeable and well-annotated dataset to combat overfitting issues. Further, a heterogeneous dataset comprising 2700 smear images is created by combining four publicly accessible benchmark datasets of ALL, AML, and healthy cases. The BSNEU-net model achieved excellent performance with 99.37% accuracy on the novel dataset and 99.44% accuracy on the heterogeneous dataset. The comparative analysis signifies the superiority of the proposed methodology with comparing schemes.
Colorectal cancer (CRC) is one of the most common causes of cancer and cancer-related mortality worldwide. Performing colon cancer screening in a timely fashion is the key to early detection. Colonoscopy is the primary modality used to diagnose colon cancer. However, the miss rate of polyps, adenomas and advanced adenomas remains significantly high. Early detection of polyps at the precancerous stage can help reduce the mortality rate and the economic burden associated with colorectal cancer. Deep learning-based computer-aided diagnosis (CADx) system may help gastroenterologists to identify polyps that may otherwise be missed, thereby improving the polyp detection rate. Additionally, CADx system could prove to be a cost-effective system that improves long-term colorectal cancer prevention. In this study, we proposed a deep learning-based architecture for automatic polyp segmentation called Transformer ResU-Net (TransResU-Net). Our proposed architecture is built upon residual blocks with ResNet-50 as the backbone and takes advantage of the transformer self-attention mechanism as well as dilated convolution(s). Our experimental results on two publicly available polyp segmentation benchmark datasets showed that TransResU-Net obtained a highly promising dice score and a real-time speed. With high efficacy in our performance metrics, we concluded that TransResU-Net could be a strong benchmark for building a real-time polyp detection system for the early diagnosis, treatment, and prevention of colorectal cancer. The source code of the proposed TransResU-Net is publicly available at https://github.com/nikhilroxtomar/TransResUNet.
Acute Lymphoblastic Leukemia (ALL) is the most prevalent form of leukemia that occurs in children. Detection of ALL through white blood cell image analysis can assist in prognosis and appropriate treatment. In this study, the author proposes an approach for classifying ALL based on white blood cell images using a Convolutional Neural Network (CNN) model called InceptionV3. The dataset used in this research consists of white blood cell images collected from patients with ALL and healthy individuals. These images were obtained from The Cancer Imaging Archive (TCIA), which is a service for storing large-scale cancer medical images available to the public. During the evaluation phase, the author used training data evaluation metrics such as accuracy and loss to measure the model's performance. The research results show that the InceptionV3 model is capable of classifying white blood cell images with a high level of accuracy. This model achieves an average ALL recognition accuracy of 0.9896 with a loss of 0.031. The use of CNN models like InceptionV3 in medical image analysis has the potential to enhance the efficiency and accuracy of image-based disease diagnosis.
A significant issue in the field of illness diagnostics is the early detection and diagnosis of leukemia, that is, the accurate distinction of malignant leukocytes with minimal costs in the early stages of the disease. Flow cytometer equipment is few, and the methods used at laboratory diagnostic centers are laborious despite the high prevalence of leukemia. The present systematic review was carried out to review the works intending to identify and categories leukemia by utilizing machine learning. It was motivated by the potential of machine learning (machine learning (ML)) in disease diagnosis. Leukemia is a blood-forming tissues cancer that affects the bone marrow and lymphatic system. It can be treated more effectively if it is detected early. This work developed a new classification model for blood microscopic pictures that distinguishes between leukemia-free and leukemia-affected images. The general proposed method in this paper consists of three main steps which are: (i) Image_Preprocessing, (ii) Feature Extraction, and (iii) Classification. An optimized CNN (OCNN) is used for classification. OCNN is utilized to detect and classify the photo as "normal" or "abnormal". Fuzzy optimization is used to optimize the hyperparameters of CNN. It is a quite beneficial to use fuzzy logic in the optimization of CNN. As illustrated from results it is shown that, with the using of OCNN classifier and after the optimization of the hyperparameters of the CNN, it achieved the best results due to the enhancement of the performance of the CNN. The OCNN has achieved 99.99% accuracy with C-NMC_Leukemia dataset.
Acute lymphoblastic leukemia (ALL) is one of the deadliest forms of leukemia due to the bone marrow producing many white blood cells (WBC). ALL is one of the most common types of cancer in children and adults. Doctors determine the treatment of leukemia according to its stages and its spread in the body. Doctors rely on analyzing blood samples under a microscope. Pathologists face challenges, such as the similarity between infected and normal WBC in the early stages. Manual diagnosis is prone to errors, differences of opinion, and the lack of experienced pathologists compared to the number of patients. Thus, computer-assisted systems play an essential role in assisting pathologists in the early detection of ALL. In this study, systems with high efficiency and high accuracy were developed to analyze the images of C-NMC 2019 and ALL-IDB2 datasets. In all proposed systems, blood micrographs were improved and then fed to the active contour method to extract WBC-only regions for further analysis by three CNN models (DenseNet121, ResNet50, and MobileNet). The first strategy for analyzing ALL images of the two datasets is the hybrid technique of CNN-RF and CNN-XGBoost. DenseNet121, ResNet50, and MobileNet models extract deep feature maps. CNN models produce high features with redundant and non-significant features. So, CNN deep feature maps were fed to the Principal Component Analysis (PCA) method to select highly representative features and sent to RF and XGBoost classifiers for classification due to the high similarity between infected and normal WBC in early stages. Thus, the strategy for analyzing ALL images using serially fused features of CNN models. The deep feature maps of DenseNet121-ResNet50, ResNet50-MobileNet, DenseNet121-MobileNet, and DenseNet121-ResNet50-MobileNet were merged and then classified by RF classifiers and XGBoost. The RF classifier with fused features for DenseNet121-ResNet50-MobileNet reached an AUC of 99.1%, accuracy of 98.8%, sensitivity of 98.45%, precision of 98.7%, and specificity of 98.85% for the C-NMC 2019 dataset. With the ALL-IDB2 dataset, hybrid systems achieved 100% results for AUC, accuracy, sensitivity, precision, and specificity.
While optical microscopy inspection of blood films and bone marrow aspirates by a hematologist is a crucial step in establishing diagnosis of acute leukemia, especially in low-resource settings where other diagnostic modalities are not available, the task remains time-consuming and prone to human inconsistencies. This has an impact especially in cases of Acute Promyelocytic Leukemia (APL) that require urgent treatment. Integration of automated computational hematopathology into clinical workflows can improve the throughput of these services and reduce cognitive human error. However, a major bottleneck in deploying such systems is a lack of sufficient cell morphological object-labels annotations to train deep learning models. We overcome this by leveraging patient diagnostic labels to train weakly-supervised models that detect different types of acute leukemia. We introduce a deep learning approach, Multiple Instance Learning for Leukocyte Identification (MILLIE), able to perform automated reliable analysis of blood films with minimal supervision. Without being trained to classify individual cells, MILLIE differentiates between acute lymphoblastic and myeloblastic leukemia in blood films. More importantly, MILLIE detects APL in blood films (AUC 0.94 ± 0.04) and in bone marrow aspirates (AUC 0.99 ± 0.01). MILLIE is a viable solution to augment the throughput of clinical pathways that require assessment of blood film microscopy.
Cloud detection is an important step in remote sensing image processing and a prerequisite for subsequent analysis and interpretation of remote sensing images. Traditional cloud detection methods are difficult to accurately detect clouds and snow with very similar features such as color and texture. In this paper, the features of cloud and snow in remote sensing images are deeply extracted, and an accurate cloud and snow detection method is proposed based on the advantages of Unet3+ network in feature fusion. Firstly, color space conversion is performed on remote sensing images, RGB images and HIS images are used as input of Unet3+ network. Resnet 50 is used to replace the Unet3+ feature extraction network to extract remote sensing image features at a deeper level, and add the Convolutional Block Attention Module in Resnet50 to improve the network’s attention to cloud and snow. Finally, the weighted cross entropy loss is constructed to solve the problem of unbalanced sample number caused by high proportion of background area in the image. The results show that the proposed method has strong adaptability and moderate computation. The mPA value, mIoU value and mPrecision value can reach 92.76%, 81.74% and 86.49%, respectively. Compared with other algorithms, the proposed method can better eliminate all kinds of interference information in remote sensing images of different landforms and accurately detect cloud and snow in images.
Stochastic-based optimization algorithms are effective approaches to addressing optimization challenges. In this article, a new optimization algorithm called the Election-Based Optimization Algorithm (EBOA) was developed that mimics the voting process to select the leader. The fundamental inspiration of EBOA was the voting process, the selection of the leader, and the impact of the public awareness level on the selection of the leader. The EBOA population is guided by the search space under the guidance of the elected leader. EBOA’s process is mathematically modeled in two phases: exploration and exploitation. The efficiency of EBOA has been investigated in solving thirty-three objective functions of a variety of unimodal, high-dimensional multimodal, fixed-dimensional multimodal, and CEC 2019 types. The implementation results of the EBOA on the objective functions show its high exploration ability in global search, its exploitation ability in local search, as well as the ability to strike the proper balance between global search and local search, which has led to the effective efficiency of the proposed EBOA approach in optimizing and providing appropriate solutions. Our analysis shows that EBOA provides an appropriate balance between exploration and exploitation and, therefore, has better and more competitive performance than the ten other algorithms to which it was compared.
Leukemia is one of the most terminal types of blood cancer, and many people suffer from it every year. White blood cells (WBCs) have a significant association with leukemia diagnosis. Research studies reported that leukemia brings changes in WBC count and morphology. WBC accurate segmentation enables to detect morphology and WBC count which consequently helps in the diagnosis and prognosis of leukemia. Manual WBC assessment methods are tedious, subjective, and less accurate. To overcome these problems, we propose a multi-scale information fusion network (MIF-Net) for WBC segmentation. MIF-Net is a shallow architecture with internal and external spatial information fusion mechanisms. In WBC images, the cytoplasm is with low contrast compared to the background, whereas nuclei shape can be complex with an indistinctive boundary for some cases, therefore accurate segmentation becomes challenging. Spatial features in the initial layers of the network include fine boundary information and MIF-Net splits and propagates this boundary information on multi-scale for external information fusion. Multi-scale information fusion in our network helps in preserving boundary information and contributes to segmentation performance improvement. MIF-Net also uses internal information fusion after intervals for feature empowerment in different stages of the network. We evaluated our network for four publicly available datasets and achieved state-of-the-art segmentation performance. In addition, the proposed architecture exhibits superior computational efficiency by using only 2.67 million trainable parameters.
Development of computer-aided cancer diagnostic tools is an active research area owing to the advancements in deep-learning domain. Such technological solutions provide affordable and easily deployable diagnostic tools. Leukaemia, or blood cancer, is one of the leading cancers causing more than 0.3 million deaths every year. In order to aid the development of such an AI-enabled tool, we collected and curated a microscopic image dataset, namely C-NMC, of more than 15000 cancer cell images at a very high resolution of B-Lineage Acute Lymphoblastic Leukaemia (B-ALL). The dataset is prepared at the subject-level and contains images of both healthy and cancer patients. So far, this is the largest (as well as curated) dataset on B-ALL cancer in the public domain. C-NMC is available at The Cancer Imaging Archive (TCIA), USA and can be helpful for the research community worldwide for the development of B-ALL cancer diagnostic tools. This dataset was utilized in an international medical imaging challenge held at ISBI 2019 conference in Venice, Italy. In this paper, we present a detailed description and challenges of this dataset. We also present benchmarking results of all the methods applied so far on this dataset.
Artificial intelligence has revolutionized medical diagnosis, particularly for cancers. Acute myeloid leukemia (AML) diagnosis is a tedious protocol that is prone to human and machine errors. In several instances, it is difficult to make an accurate final decision even after careful examination by an experienced pathologist. However, computer-aided diagnosis (CAD) can help reduce the errors and time associated with AML diagnosis. White Blood Cells (WBC) detection is a critical step in AML diagnosis, and deep learning is considered a state-of-the-art approach for WBC detection. However, the accuracy of WBC detection is strongly associated with the quality of the extracted features used in training the pixel-wise classification models. CAD depends on studying the different patterns of changes associated with WBC counts and features. In this study, a new hybrid feature extraction method was developed using image processing and deep learning methods. The proposed method consists of two steps: 1) a region of interest (ROI) is extracted using the CMYK-moment localization method and 2) deep learning-based features are extracted using a CNN-based feature fusion method. Several classification algorithms are used to evaluate the significance of the extracted features. The proposed feature extraction method was evaluated using an external dataset and benchmarked against other feature extraction methods. The proposed method achieved excellent performance, generalization, and stability using all the classifiers, with overall classification accuracies of 97.57% and 96.41% using the primary and secondary datasets, respectively. This method has opened a new alternative to improve the detection of WBCs, which could lead to a better diagnosis of AML.
Acute Lymphoblastic Leukemia (ALL) is a blood cell cancer characterized by the presence of excess immature lymphocytes., Even though automation in ALL prognosis is essential for cancer diagnosis, it remains a challenge due to the morphological correlation between malignant and normal cells. The traditional ALL classification strategy demands that experienced pathologists read cell images carefully, which is arduous, time-consuming, and often hampered by interobserver variation. This article has automated the ALL recognition task by employing deep Convolutional Neural Networks (CNNs). The weighted ensemble of deep CNNs is explored to recommend a better ALL cell classifier. The weights are estimated from ensemble candidates’ corresponding metrics, such as F1-score, area under the curve (AUC), and kappa values. Various data augmentations and pre-processing are incorporated to achieve a better generalization of the network. Our proposed model was trained and evaluated utilizing the C-NMC-2019 ALL dataset. The proposed weighted ensemble model has outputted a weighted F1-score of 89.7%, a balanced accuracy of 88.3%, and an AUC of 0.948 in the preliminary test set. The qualitative results displaying the gradient class activation maps confirm that the introduced model has a concentrated learned region. In contrast, the ensemble candidate models, such as Xception, VGG-16, DenseNet-121, MobileNet, and InceptionResNet-V2, separately exhibit coarse and scatter learned areas in most cases. Since the proposed ensemble yields a better result for the aimed task, it can support clinical decisions to detect ALL patients in an early stage.
In the human body, one of the fatal diseases of white blood cells is leukemia that affects the bone marrow as well as blood. In this work, the leukemia image is pre-processed with a data augmentation model in which the noise and over-fitting are eliminated and the data augmentation increases the size of the datasets. We used Hybrid Convolutional Neural Network with Interactive Autodidactic School (HCNN-IAS) algorithm, which performs feature extraction, fusing, and classification operations. At first, the local and global features are extracted from the leukemia image. After that, the self-attention in CNN combines both local and global features. Finally, the HCNN-IAS algorithm effectively categorizes different classes of leukemia such as healthy, Acute Lymphocytic Leukemia (ALL), Acute Myeloid Leukemia (AML), Chronic myeloid leukemia (CML), and Chronic Lymphocytic Leukemia(CLL). The proposed model is implemented in an Internet of Medical Things (IoMT) platform to offer the appropriate treatment to the patients within the comfort of their homes. The images of the dataset were acquired from the ASH image bank. The experimental result demonstrates better and higher classification accuracy in terms of leukemia detection with reasonable accuracy, precision and recall rates of about 99%.
Akin the introduction of tyrosine kinase inhibitors to Philadelphia chromosome-positive acute lymphoblastic leukemia (ALL), pediatric-based asparaginase-heavy approaches have revolutionized the treatment of young adults with the Philadelphia chromosome-negative subset the past decades. Once again, we are approaching a new era. An era of precision medicine with immunotherapy and other molecularly targeted treatments that offers unique opportunities to customize treatment intensity with or without hematopoietic stem cell transplantation, reduce the burden of toxicities, and combat persistent residual disease. Recently approved agents for refractory/relapsed B-cell precursor ALL include the chimeric antigen receptor-modified (CAR) T-cells, the anti-CD22 monoclonal antibody-drug conjugate, inotuzumab ozogamicin, and the bispecific anti-CD19 T-cell engager, blinatumomab. These agents are expected to move widely into the frontline setting along with the proteasome inhibitors, bortezomib and carfilzomib, as well as tyrosine kinase inhibitors for Philadelphia-like rearrangements that are especially frequent among young adults. To this add the BH3 mimetics, venetoclax and navitoclax, which are being widely explored in refractory/relapsed as well as frontline settings for B- and T-cell ALL. The promising anti-CD38 monoclonal antibody, daratumumab, is entering the scene of refractory/relapsed T-ALL, whereas the old purine analogue, nelarabine, is being evaluated in a new upfront setting. This review focuses on two main questions: How do we optimize frontline as well as salvage ALL treatment of young adults in the 2020s? Not least, how do we address the current burden of serious toxicities unique to young adults?
Leukocytes, produced in the bone marrow, make up around one percent of all blood cells. Uncontrolled growth of these white blood cells leads to the birth of blood cancer. Out of the three different types of cancers, the proposed study provides a robust mechanism for the classification of Acute Lymphoblastic Leukemia (ALL) and Multiple Myeloma (MM) using the SN-AM dataset. Acute lymphoblastic leukemia (ALL) is a type of cancer where the bone marrow forms too many lymphocytes. On the other hand, Multiple myeloma (MM), a different kind of cancer, causes cancer cells to accumulate in the bone marrow rather than releasing them into the bloodstream. Therefore, they crowd out and prevent the production of healthy blood cells. Conventionally, the process was carried out manually by a skilled professional in a considerable amount of time. The proposed model eradicates the probability of errors in the manual process by employing deep learning techniques, namely convolutional neural networks. The model, trained on cells’ images, first pre-processes the images and extracts the best features. This is followed by training the model with the optimized Dense Convolutional neural network framework (termed DCNN here) and finally predicting the type of cancer present in the cells. The model was able to reproduce all the measurements correctly while it recollected the samples exactly 94 times out of 100. The overall accuracy was recorded to be 97.2%, which is better than the conventional machine learning methods like Support Vector Machine (SVMs), Decision Trees, Random Forests, Naive Bayes, etc. This study indicates that the DCNN model’s performance is close to that of the established CNN architectures with far fewer parameters and computation time tested on the retrieved dataset. Thus, the model can be used effectively as a tool for determining the type of cancer in the bone marrow.
Computer-aided diagnosis system is becoming a more and more important tool in clinical treatment, which can provide a verification of the doctors’ decisions. In this paper, we proposed a novel abnormal brain detection method for magnetic resonance image. Firstly, a pre-trained AlexNet was modified with batch normalization layers and trained on our brain images. Then, the last several layers were replaced with an extreme learning machine. A searching method was proposed to find the best number of layers to be replaced. Finally, the extreme learning machine was optimized by chaotic bat algorithm to obtain better classification performance. Experiment results based on 5 × hold-out validation revealed that our method achieved state-of-the-art performance.
The works presented in this paper addresses the robust population-based global optimization that is influenced by the simplicity and efficiency principles introduced in two new generation optimization algorithms. Galactic Swarm Optimization is inspired by the motion of stars, galaxies, and superclusters of galaxies under the influence of gravity. It acts well as a global controller of the whole optimization process by employing multiple flexible cycles of exploration and exploitation phases to find new, better solutions. However, the optimization process still suffers poverty in the exploitation phase, which is improved in this work by its hybridization with our evolution version of the Whale Optimization Algorithm. Concretely, the exploitation phase of Galactic Swarm Optimization is replaced by our Evolution Whale Optimization Algorithm to avoid early convergence. The Whale Optimization Algorithm mimics the unusual social behaviors and the hunting activities of humpback whales. However, it is not optimized for global optimization when the number of dimensions is increased. Hence its is evolved in our works by Levy-Flight trajectory for faster local search with adaptive step lengths and two-point crossover operator to reduce bias in the offspring creation procedure. The achieved results through extensive and careful experiments showed that our hybridization and evolution enhancements bring outstanding performance in terms of accuracy, convergence speed, and stability.
Methods:
We evaluate the effectiveness of very deep convolutional neural networks at the task of expert-level lung nodule malignancy classification. Using the state-of-the-art ResNet architecture as our basis, we explore the effect of curriculum learning, transfer learning, and varying network depth on the accuracy of malignancy classification.
Results:
Due to a lack of public datasets with standardized problem definitions and train/test splits, studies in this area tend to not compare directly against other existing work. This makes it hard to know the relative improvement in the new solution. In contrast, we directly compare our system against two state-of-the-art deep learning systems for nodule classification on the LIDC/IDRI dataset using the same experimental setup and data set. The results show that our system achieves the highest performance in terms of all metrics measured including sensitivity, specificity, precision, AUROC, and accuracy.
Conclusions:
The proposed method of combining deep residual learning, curriculum learning, and transfer learning translates to high nodule classification accuracy. This reveals a promising new direction for effective pulmonary nodule CAD systems that mirrors the success of recent deep learning advances in other image-based application domains.
Integrating artificial intelligence with food category recognition has been a field of interest for research for the past few decades. It is potentially one of the next steps in revolutionizing human interaction with food. The modern advent of big data and the development of data-oriented fields like deep learning have provided advancements in food category recognition. With increasing computational power and ever-larger food datasets, the approach's potential has yet to be realized. This survey provides an overview of methods that can be applied to various food category recognition tasks, including detecting type, ingredients, quality, and quantity. We survey the core components for constructing a machine learning system for food category recognition, including datasets, data augmentation, hand-crafted feature extraction, and machine learning algorithms. We place a particular focus on the field of deep learning, including the utilization of convolutional neural networks, transfer learning, and semi-supervised learning. We provide an overview of relevant studies to promote further developments in food category recognition for research and industrial applications.
Bone Marrow Cancer is a type of cancer that develops in the stem cells of the bone marrow that are responsible for blood formation. AML(Acute Myeloid Leukaemia) and MM(Multiple Myeloma) are both types of malignancy that can affect bone marrow. Conventionally, the process was carried out manually by a skilled professional in a considerable amount of time. Hence, in order to provide accurate treatment, a better diagnosis technique is required. In this regard, the suggested study has improved a unique technique, specifically classification and segmentation, allowing for the identification of more intricate disorders. The study's findings on the accuracy of cancer cell identification, as well as the decrease in the possibility of false alarm rates, continue to fall on the negative side of the outcome spectrum, according to the researchers. The hybrid approach has been proposed in this article that includes deep convolutional neural networks, hyperparameter sets utilising adaptive multi-objective CAT algorithms and other image processing techniques. To train the suggested model, it is critical to use previously processed cell images. The suggested model is then used to train an Optimized Convolutional Neural Network (OCNN), which is then used to determine the type of tumour identified in the bone area. The SN-AM datasets were thoroughly examined, and a number of presentation measures, such as accuracy (recall), F1-score, and specificity (specificity), were generated and analysed. When it came to predicting different types of cancer, researchers observed that they had an overall accuracy of 99.45 percent. In terms of recognizing the sorts of cancer cells, the proposed model surpassed all existing learning models.
Farmland segmentation scenario from remote sensing images plays an important role in crop growth monitoring, precision agriculture and intelligent agriculture. To achieve high precision segmentation of farmland boundary, a Multi-task Deformable UNet combined Enhanced network (MDE-UNet) is proposed for farmland boundary segmentation. The network consists of two parts: a Multi-task Deformable UNet (MD-UNet) segmentation module with Deformable UNet (D-UNet) as the basic network and an enhancement module with a lightweight UNet improved by residual attention. In the MD-UNet segmentation module, three branches are used for precise segmentation of deterministic, fuzzy, and raw boundary, respectively. In the enhancement module, an improved lightweight UNet is designed, which can enhance the feature extraction ability of the MD-UNet segmentation module and further improve the segmentation accuracy. The accuracy and mIoU in the GF-2 farmland segmentation test dataset can reach 96.41% and 91.29% using the proposed model, respectively. The MDE-UNet method outperforms other representative deep learning methods such as DeepLab v3+, FCN-8s, SegFormer, and UTNet, and has potential for practical applications of farmland boundary segmentation.
Analysis of gene expression data is crucial for disease prognosis and diagnosis. Gene expression data has high redundancy and noise that brings challenges in extracting disease information. Over the past decade, several conventional machine learning and deep learning models have been developed for classification of diseases using gene expressions. In recent years, vision transformer networks have shown promising performance in many fields due to their powerful attention mechanism that provides a better insight into the data characteristics. However, these network models have not been explored for gene expression analysis. In this paper, a method for classifying cancerous gene expression is presented that uses a Vision transformer. The proposed method first performs dimensionality reduction using a stacked autoencoder followed by an Improved DeepInsight algorithm that converts the data into image format. The data is then fed to the vision transformer for building the classification model. Performance of the proposed classification model is evaluated on ten benchmark datasets having binary classes or multiple classes. Its performance is also compared with nine existing classification models. The experimental results demonstrate that the proposed model outperforms existing methods. The t-SNE plots demonstrate the distinctive feature learning property of the model.
For the increasingly serious phenomenon of electricity theft, many researchers are trying to detect it. Traditional detection methods rely on physical inspection, which has low detection efficiency and high cost. However, the rapid application of advanced metering infrastructure (AMI) makes it possible to detect electricity theft via smart meters. In this work, we propose a hybrid improved wide and deep convolutional neural networks (CNN) method to address electricity theft detection on an imbalanced realistic dataset that presents a daily electricity consumption provided by State Grid Corporation of China. Our key contribution is the introduction of a channel-dimensional adaptive attention module concatenated with dilated convolutions. Moreover, we propose focal loss to solve the data imbalance problem. Experimental results demonstrate that compared to other state-of-the-art deep learning classifiers, our improved method has better performance (with 97.08% of [email protected] and 83.61% of AUC), which verifies the effectiveness of the proposed method.
Context:
The ordinary morphologic diagnosis of acute lymphoblastic leukemia (ALL) by a pathologist depends on examining the patient's peripheral blood (PB) together with the bone marrow (BM) blood films. However, this manual aspect of diagnosis is susceptible to discrepancies. We now have a newly introduced technology that allows us to overcome individual variation in the diagnosis of ALL, so-named machine learning, which depends on a complex, preprogrammed convolutional networks matrix.
Objectives:
Challenging machine-aided systems that utilize microscopic blood film images to recognize and diagnose ALL based on a preprogrammed deep convolutional neural network (CNN), i.e., machine learning algorithms.
Material/method:
We collected a dataset of images composed of PB and BM smear images of two classes: ALL and normal control blood. We analyzed 192 samples of digital images: 96 images of patients with ALL and 96 images of healthy normal controls. For each smear sample, we collected the results of clinical data (clinical history and examination) and laboratory data (morphological, cytochemical, and immunophenotyping assessment). We challenged seven types of CNN models to diagnose ALL: AlexNet, VGG16, VGG19, GoogLeNet, ResNet50, ResNet101, and Inception-v3.
Results:
Comparing the ability of seven models to diagnose ALL revealed that the AlexNet had the lowest accuracy of 95.51%, followed by VGG16 (92.13%) and VGG19 (93.83%). Inception-v3 had the highest accuracy (99.98%) and was able to detect almost all ALL cases.
Conclusions:
The statistical measures of the Inception-v3 performance revealed promising results in detecting ALL cases: the sensitivity, specificity, and accuracy of Inception-v3 reached 99.98% for detection of ALL.
Cyclin-dependent kinase 9 (CDK9) plays a vital role in controlling cell transcription and has been an attractive target for cancer treatment. Herein, ten predictive models derived from 1330 unique molecules against CDK9 were constructed based on molecular fingerprints and graphs using two conventional machine learning and four deep learning methods. The evaluation results showed that FP-GNN deep learning architecture performed best for CDK9 inhibitors prediction with the highest BA and F1 values of 0.681 and 0.912 for testing set. We then performed virtual screening to identify new CDK9 inhibitors by incorporating the optimal established predictive model and molecular docking. Five compounds were identified to show broad anticancer activity against various cancer cell lines through bioassays. For example, C9 exhibited antiproliferative activity against HeLa, MOLM-13 and MDA-MB-231 with IC50 of 2.53, 3.92 and 11.65 μM. Kinase inhibition assay results demonstrated that these compounds displayed submicromolar (214∼504 nM) inhibitory activities against CDK9. Further cellular mechanism evaluation revealed that C9 suppressed the activity of CDK9 and interfered with the expression of Mcl-1 and cleaved PARP in MOLM-13 cells, resulting in the induction of cellular apoptosis. In addition, C9 displayed a good stability in rat liver microsomes, artificial gastrointestinal fluid and plasm. An online platform (called DEEPCDK9Pred) was developed based on the FP-GNN models to predict or design new CDK9 inhibitors. Collectively, our findings demonstrated that FP-GNN algorithm can achieve accurate prediction of CDK9 inhibitors and the subsequent discovery of C9 as a new potential CDK9 inhibitor deserves further structural modification for the treatment of leukemia.
Acute Lymphoblastic Leukemia (ALL) is cancer in which bone marrow overproduces undeveloped lymphocytes. Over 6500 cases of ALL are diagnosed every year in the United States in both adults and children, accounting for around 25% of pediatric cancers, and the trend continues to rise. With the advancements of AI and big data analytics, early diagnosis of ALL can be used to aid the clinical decisions of physicians and radiologists. This research proposes a deep neural network-based (ALNet) model that employs depth-wise convolution with different dilation rates to classify microscopic white blood cell images. Specifically, the cluster layers encompass convolution and max-pooling followed by a normalization process that provides enriched structural and contextual details to extract robust local and global features from the microscopic images for the accurate prediction of ALL. The performance of the model was compared with various pre-trained models, including VGG16, ResNet-50, GoogleNet, and AlexNet, based on precision, recall, accuracy, F1 score, loss accuracy, and receiver operating characteristic (ROC) curves. Experimental results showed that the proposed ALNet model yielded the highest classification accuracy of 91.13% and an F1 score of 0.96 with less computational complexity. ALNet demonstrated promising ALL categorization and outperformed the other pre-trained models.
An accurate early-stage leukemia diagnosis plays a critical role in treating and saving patients’ lives. The two primary forms of leukemia are acute and chronic leukemia, which is subdivided into myeloid and lymphoid leukemia. Deep learning models have been increasingly used in computer-aided medical diagnosis (CAD) systems developed to detect leukemia. This article assesses the impact of widely applied techniques, mainly data augmentation and multilevel and ensemble configurations, in deep learning-based CAD systems. Our assessment included five scenarios: three binary classification problems and two multiclass classification problems. The evaluation was performed using 3,536 images from 18 datasets, and it was possible to conclude that data augmentation techniques improve the performance of convolutional neural networks (CNNs). Furthermore, there is an improvement in the classification results using a combination of CNNs. For the binary problems, the performance of the ensemble configuration was superior to that of the multilevel configuration. However, the results were statistically similar in multiclass scenarios. The results were promising, with accuracies of 94.73% and 94.59% obtained using multilevel and ensemble configurations in a scenario with four classes. The combination of methods helps to reduce the error or variance of the predictions, which improves the accuracy of the used deep learning-based model.
Background:
Classification of acute myeloid leukemia (AML) relies on manual analysis of bone marrow or peripheral blood smear images. We aimed to construct a machine learning model for automatic classification of AML-M1 and M2 subtypes in bone marrow smear images.
Methods:
Bone marrow smear images of AML patients were extracted from the Cancer Imaging Archive (TCIA) open database. Classification criteria of AML subtypes were based on the French-American-British (FAB) classification system. Random forest method and broad learning system (BLS) were used to develop the classification model. Morphological features, radiomics features, and clinical features were extracted. The performance of the classification model was evaluated by calculating accuracy, precision, recall, F1-score, and area under the curve (AUC). A total of 50 bone marrow smear images (AML-M1, 31 cases; AML-M2, 19 cases) with 500 slices were included in this study.
Results:
A total of 43 morphological features, 276 radiomics features, and 1 clinical feature were extracted. Finally, 9 variables including 2 morphological features, 6 radiomics features, and 1 clinical feature were selected into the classification model. The best classification performance was observed in the random forest model with 9 variables, with the average accuracy, AUC, F1-score, recall, and precision of the model being 0.998 ± 0.003, 0.998 ± 0.004, 0.998 ± 0.004, 0.996 ± 0.009, and 1 ± 0, respectively.
Conclusion:
The random forest model performed well for the classification of AML-M1 and M2, which may provide a tool for clinicians to classify AML-M1 and M2.
Chronic myeloid leukemia (CML) is a clonal proliferative disorder of granulocytic lineage, and morphological evaluation is the first step for a definite diagnosis. This study developed a conditional generative adversarial network (cGAN) based model, CMLcGAN, to segment megakaryocytes (MKs) from myeloid cells in bone marrow biopsies. After the segmentation, the statistical characteristics of two types of cells were extracted and compared between patients and controls. At the segmentation phase, the CMLcGAN was evaluated on 517 images (512 × 512), and achieved a mean pixel accuracy of 95.1%, a mean Intersection over Union (IoU) of 71.2%, and a mean Dice coefficient of 81.8%. In addition, the CMLcGAN was compared with seven other available deep learning-based segmentation models, and achieved a better segmentation performance. At the clinical validation phase, a series of 7-dimensional statistical features from various cells were extracted. Using the T-test, 5-dimensional features were selected as the clinical prediction feature set. Finally, the model iterated 100 times using 3-fold cross-validation on whole slide images (58 CML cases and 31 normal cases), and the final best AUC was 84.93%. In conclusion, a CMLcGAN model was established for multi-class segmentation of bone marrow cells and performed better than other deep learning-based segmentation models.
Crack detection is an indispensable premise of road maintenance, which can provide early warning information for many road damages and save repair costs to a large extent. Because of the security and convenience, many image processing technique (IPT) based crack detection methods have been proposed, but their performances often cannot meet the requirements of practical applications because of the complex texture structure and seriously imbalanced categories. To address the aforementioned problem, we present an external attention based TransUNet for crack detection. Specifically, we tackle the TransUNet as the backbone of our detection framework, which can propagate the detailed texture information from shallow layers to corresponding deep layers through skip connections. Besides, the Transformer Block equipped in the second last convolution layer of the encoding component can explicitly model the long-range dependency of different regions in an image, which improves the structural representation ability of the framework and hence alleviates the interference from shadow, noise, and other negative factors. In addition, the External Attention Block equipped in the last convolution layer of the encoding component can effectively exploit the dependency of crack regions among different images, and further enhance the robustness of the framework. Finally, combined with the Focal Loss, the proposed label expansion strategy can further alleviate the category imbalance problem through transforming semantic categories of non-crack pixels distributed in the neighbors of corresponding crack pixels.
Mutations characterize diverse human cancers; there is a positive correlation between elevated mutation frequency and tumor progression. One exception is acute myeloid leukemia (AML), which has few clonal single nucleotide mutations. We used highly sensitive and accurate Duplex Sequencing (DS) to show now that AML, in addition, has an extensive repertoire of variants with low allele frequencies, <1%, which is below the accurate detection limit of most other sequencing methodologies. The subclonal variants are unique to each individual and change in composition, frequency, and sequence context from diagnosis to relapse. Their functional significance is apparent by the observation that many are known variants and cluster within functionally important protein domains. Subclones provide a reservoir of variants that could expand and contribute to the development of drug resistance and relapse. In accord, we accurately identified subclonal variants in AML driver genes NRAS and RUNX1 at allele frequencies between 0.1-0.3% at diagnosis, which expanded to comprise a major fraction (14-53%) of the blast population at relapse. Early and accurate detection of subclonal variants with low allele frequency thus offers the opportunity for early intervention, prior to detection of clinical relapse, to improve disease outcome and enhance patient survival.
Microscopic analysis of blood-cells is an essential and vital task for the early diagnosis of life-threatening hematological disorders like blood cancer (leukemia). We have presented an effective and computationally efficient approach for automatically detecting and classifying Acute Lymphoblastic Leukemia (ALL) and Acute Myeloid Leukemia (AML). Currently, transfer learning has succeeded as a preferred approach in medical image analysis since it achieves excellent performance in a small database. This paper proposes a lightweight transfer-learning-based feature extraction followed by Support Vector Machine (SVM)-based classification technique for efficient ALL and AML detection. It yields a faster and more efficient system due to the depth-wise separable convolution, tunable multiplier, and inverted residual bottleneck structure. Moreover, the SVM-based classification improves the overall performance by optimizing the hyperplane location. Furthermore, the experimental results signify that our proposed system gains superior performance than others in all these three publicly available standard ALLIDB1, ALLIDB2, and ASH databases.
COVID-19 pneumonia started in December 2019 and caused large casualties and huge economic losses. In this study, we intended to develop a computer-aided diagnosis system based on artificial intelligence to automatically identify the COVID-19 in chest computed tomography images. We utilized transfer learning to obtain the image-level representation (ILR) based on the backbone deep convolutional neural network. Then, a novel neighboring aware representation (NAR) was proposed to exploit the neighboring relationships between the ILR vectors. To obtain the neighboring information in the feature space of the ILRs, an ILR graph was generated based on the k-nearest neighbors algorithm, in which the ILRs were linked with their k-nearest neighboring ILRs. Afterward, the NARs were computed by the fusion of the ILRs and the graph. On the basis of this representation, a novel end-to-end COVID-19 classification architecture called neighboring aware graph neural network (NAGNN) was proposed. The private and public data sets were used for evaluation in the experiments. Results revealed that our NAGNN outperformed all the 10 state-of-the-art methods in terms of generalization ability. Therefore, the proposed NAGNN is effective in detecting COVID-19, which can be used in clinical diagnosis.
Background:
Conventional identification of blood disorders based on visual inspection of blood smears through microscope is time consuming, error-prone and is limited by hematologist's physical acuity. Therefore, an automated optical image processing system is required to support the clinical decision-making.
Materials and methods:
Blood smear slides (n = 250) were prepared from clinical samples, imaged and analyzed in Jimma Medical Center, Hematology department. Samples were collected, analyzed and preserved from out and in-patients. The system was able to categorize four common types of leukemia's such as acute and chronic myeloid leukemia; and acute and chronic lymphoblastic leukemia, through a robust image segmentation protocol, followed by classification using the support vector machine.
Results:
The system was able to classify leukemia types with an accuracy, sensitivity, specificity of 97.69%, 97.86% and 100%, respectively for the test datasets, and 97.5%, 98.55% and 100%, respectively, for the validation datasets. In addition, the system also showed an accuracy of 94.75% for the WBC counts that include both lymphocytes and monocytes. The computer-assisted diagnosis system took less than one minute for processing and assigning the leukemia types, compared to an average period of 30 minutes by unassisted manual approaches. Moreover, the automated system complements the healthcare workers' in their efforts, by improving the accuracy rates in diagnosis from ∼70% to over 97%.
Conclusion:
Importantly, our module is designed to assist the healthcare facilities in the rural areas of sub-Saharan Africa, equipped with fewer experienced medical experts, especially in screening patients for blood associated diseases including leukemia.
Automated and accurate diagnosis of Acute Lymphoblastic Leukemia (ALL), blood cancer, is a challenging task. Nowadays, Convolutional Neural Networks (CNNs) have become a preferred approach for medical image analysis. However, for achieving excellent performance, classical CNNs usually require huge databases for proper training. This paper proposes an efficient deep CNNs framework to mitigate this issue and yield more accurate ALL detection. The salient features: depthwise separable convolutions, linear bottleneck architecture, inverted residual, and skip connections make it a faster and preferred approach. In this proposed method, a novel probability-based weight factor is suggested, which has a significant role in efficiently hybridizing MobilenetV2 and ResNet18 with preserving the benefits of both approaches. Its performance is validated using public benchmark datasets: ALLIDB1 and ALLIDB2. The experimental results display that the proposed approach yields the best accuracy (with 70% training and 30% testing) 99.39% and 97.18% in ALLIDB1 and ALLIDB2 datasets, respectively. Similarly, it also achieves the best accuracy (with 50% training and 50% testing) 97.92% and 96.00% in ALLIDB1 and ALLIDB2 datasets, respectively. Moreover, it also achieves the best performance compared to the recent transfer learning-based techniques in both the datasets, in terms of sensitivity, specificity, accuracy, precision, F1 score, and receiver operating characteristic (ROC) in most of the cases.
Purpose
Leukaemia is diagnosed conventionally by observing the peripheral blood and bone marrow smear using a microscope and with the help of advanced laboratory tests. Image processing-based methods, which are simple, fast, and cheap, can be used to detect and classify leukemic cells by processing and analysing images of microscopic smear. The proposed study aims to classify Acute Lymphoblastic Leukaemia (ALL) by Deep Learning (DL) based techniques.
Procedures
The study used Deep Convolutional Neural Networks (DNNs) to classify ALL according to WHO classification scheme without using any image segmentation and feature extraction that involves intense computations. Images from an online image bank of American Society of Haematology (ASH) were used for the classification.
Findings
A classification accuracy of 94.12% is achieved by the study in isolating the B-cell and T-cell ALL images using a pretrained CNN AlexNet as well as LeukNet, a custom-made deep learning network designed by the proposed work. The study also compared the classification performances using three different training algorithms.
Conclusions
The paper detailed the use of DNNs to classify ALL, without using any image segmentation and feature extraction techniques. Classification of ALL into subtypes according to the WHO classification scheme using image processing techniques is not available in literature to the best of the knowledge of the authors. The present study considered the classification of ALL only, and detection of other types of leukemic images can be attempted in future research.
Acute Lymphoblastic Leukemia (ALL) is a cancer of the blood cells which is characterized by a large number of immature lymphocytes, known as blast cells (myeloblasts). To aid the ALL diagnosis, we propose to automate the blast cell detection using Artificial Intelligence (AI). Our automation system incorporates an object detection method that predicts leukemic cells from microscopic blood smear images. We have implemented version 4 of the You Only Look Once (YOLOv4) algorithm for both cell detection and cell classification. As such, the classification was set up as a binary problem, where each cell was labeled as either blast cells (ALL) or healthy cells (HEM). The Object Detection algorithm was trained and tested with images from the ALL_IDB1 and C_NMC_2019 dataset. The mAP (Mean Average Precision) was 96.06 % for the ALL-IDB1 dataset and 98.7 % for the C_NMC_2019 dataset. Both models were trained with Google Colaboratory using a Nvidia Tesla P-100 GPU. This proposed blast cell detection algorithm might be used as an adjunct tool during pre-screening where it can help to detect Leukemia based on microscopic blood smear images.
Background and objectives
Morphological differentiation among blasts circulating in blood in acute leukaemia is challenging. Artificial intelligence decision support systems hold substantial promise as part of clinical practise in detecting haematological malignancy. This study aims to develop a deep learning-based system to predict the diagnosis of acute leukaemia using blood cell images.
Methods
A set of 731 blood smears containing 16,450 single-cell images was analysed from 100 healthy controls, 191 patients with viral infections and 148 with acute leukaemia. Training and testing sets were arranged with 85% and 15% of these smears, respectively. To find the best architecture for acute leukaemia classification VGG16, ResNet101, DenseNet121 and SENet154 were evaluated. Fine-tuning was implemented to these pre-trained CNNs to adapt their layers to our data. Once the best architecture was chosen, a system with two modules working sequentially was configured (ALNet). The first module recognised abnormal promyelocytes among other mononuclear blood cell images, such as lymphocytes, monocytes, reactive lymphocytes and blasts. The second distinguished if blasts were myeloid or lymphoid lineage. The final strategy was to predict patients’ initial diagnosis of acute leukaemia lineage using the blood smear review. ALNet was assessed with smears of the testing set.
Results
ALNet provided the correct diagnostic prediction of all patients with promyelocytic and myeloid leukaemia. Sensitivity, specificity and precision values of 100%, 92.3% and 93.7%, respectively, were obtained for myeloid leukaemia. Regarding lymphoid leukaemia, a sensitivity of 89% and specificity and precision values of 100% were obtained.
Conclusions
ALNet is a predictive model designed with two serially connected convolutional networks. It is proposed to assist clinical pathologists in the diagnosis of acute leukaemia during the blood smear review. It has been proved to distinguish neoplastic (leukaemia) and non-neoplastic (infections) diseases, as well as recognise the leukaemia lineage.
This paper presents a novel meta-heuristic algorithm named Chameleon Swarm Algorithm (CSA) for solving global numerical optimization problems. The base inspiration for CSA is the dynamic behavior of chameleons when navigating and hunting for food sources on trees, deserts and near swamps. This algorithm mathematically models and implements the behavioral steps of chameleons in their search for food, including their behavior in rotating their eyes to a nearly 360°scope of vision to locate prey and grab prey using their sticky tongues that launch at high speed. These foraging mechanisms practiced by chameleons eventually lead to feasible solutions when applied to address optimization problems. The stability of the proposed algorithm was assessed on sixty-seven benchmark test functions and the performance was examined using several evaluation measures. These test functions involve unimodal, multimodal, hybrid and composition functions with different levels of complexity. An extensive comparative study was conducted to demonstrate the efficacy of CSA over other meta-heuristic algorithms in terms of optimization accuracy. The applicability of the proposed algorithm in reliably addressing real-world problems was demonstrated in solving five constrained and computationally expensive engineering design problems. The overall results of CSA show that it offered a favorable global or near global solution and better performance compared to other meta-heuristics.
Purpose
To evaluate the efficacy of diagnosis systems based upon instance segmentation with convolutional neural networks (CNNs) for diagnosing acute promyelocytic leukemia (APL) in bone marrow smear images.
Materials and methods
A self-established dataset was used in this study that was exempted from review by the institution review board, which consisted of 13,504 bone marrow smear images. One subset of the dataset with 12,215 labeled images was split into training (80%) and validation (20%), another with 1289 labeled images was used to test, in which each test entry consists of about 130 images. An instance segmentation method named Mask R-CNN was used to detect and classify the nucleated cells. Here, we train a trained neural network from scratch; for comparison, we also use a network pre-trained on MS COCO (common objects in context, a data set provided by Microsoft which can be used for image recognition, the images in MS coco dataset are divided into training, validation and test sets) and fine-tuned with our dataset and both were trained with same data augmentation scheme. Diagnosis systems based on trained models and “FAB Classification” (French–American–British classification systems, a series of diagnostic criteria for acute leukemia, which was first proposed in 1976) were developed for diagnosing the test entry as APL or as not. Average precision (AP) and average recall (AR) were used to evaluate model performance.
Results
The best-performing model had an average precision of 62.5%, which was the augmented pre-trained Mask R-CNN with average recall 84.1%. The average precision of the pre-trained model was greater than that of the model trained from scratch (P < 0.05). Augmenting the dataset further increased accuracy (P < 0 0.03).
Conclusion
Deep learning technology such as instance segmentation with Mask R-CNN may accurately diagnose APL in bone marrow smear images with an average precision of 62.5% when 0.5 as IoU thresholds. A data augmentation and pre-trained approach further improved accuracy.
The plethora of ways of data representation and their applications to system modeling is inherently associated with dimensionality reduction. In a nutshell, the result of dimensionality reduction should support efficient ways of constructing ensuing models (classifiers, predictors) as well as an interpretation of the data themselves. Furthermore, there should be a suitable measure quantifying the quality of data positioned in the reduced space. We advocate that what makes the reduced data interpretable, goes hand in hand with revealing a logic fabric of the data, suppressing redundancy, and finally arriving at a logic description of data. The anticipation is that the reduced data can be described in the form of logic expressions formed over the original highly dimensional data. Evidently, having these above stated points in mind, the aim of this study is two-fold: (i) to develop a logic-oriented data representation, and (ii) to quantify the quality of results of dimensionality reduction by incorporating a facet of information granularity. In other words, we argue that the result of dimensionality reduction gives rise to information granules whose level of granularity associates with the quality of processing completed by the autoencoder. In light of the recent surge of architectures of deep learning, the study is focused on the construction and analysis of logic-oriented autoencoders. We propose a two-level architecture composed of the logic-oriented processing units (and processing carried out at the first layer of the autoencoder) followed by the or processing completed at the second layer. As data representation provided by the autoencoder is not ideal, we augment the original architecture by granular parameters which give rise to granular logic-oriented autoencoders. A suite of experiments is also reported.
This paper presents a novel CNN-RNN based approach, which exploits multiple CNN features for dimensional emotion recognition in-the-wild, utilizing the One-Minute Gradual-Emotion (OMG-Emotion) dataset. Our approach includes first pre-training with the relevant and large in size, Aff-Wild and Aff-Wild2 emotion databases. Low-, mid- and high-level features are extracted from the trained CNN component and are exploited by RNN subnets in a multi-task framework. Their outputs constitute an intermediate level prediction; final estimates are obtained as the mean or median values of these predictions. Fusion of the networks is also examined for boosting the obtained performance, at Decision-, or at Model-level; in the latter case a RNN was used for the fusion. Our approach, although using only the visual modality, outperformed state-of-the-art methods that utilized audio and visual modalities. Some of our developments have been submitted to the OMG-Emotion Challenge, ranking second among the technologies which used only visual information for valence estimation; ranking third overall. Through extensive experimentation, we further show that arousal estimation is greatly improved when low-level features are combined with high-level ones.
Although end-to-end neural text-to-speech (TTS) methods (such as Tacotron2) are proposed and achieve state-of-theart performance, they still suffer from two problems: 1) low efficiency during training and inference; 2) hard to model long dependency using current recurrent neural networks (RNNs). Inspired by the success of Transformer network in neural machine translation (NMT), in this paper, we introduce and adapt the multi-head attention mechanism to replace the RNN structures and also the original attention mechanism in Tacotron2. With the help of multi-head self-attention, the hidden states in the encoder and decoder are constructed in parallel, which improves training efficiency. Meanwhile, any two inputs at different times are connected directly by a self-attention mechanism, which solves the long range dependency problem effectively. Using phoneme sequences as input, our Transformer TTS network generates mel spectrograms, followed by a WaveNet vocoder to output the final audio results. Experiments are conducted to test the efficiency and performance of our new network. For the efficiency, our Transformer TTS network can speed up the training about 4.25 times faster compared with Tacotron2. For the performance, rigorous human tests show that our proposed model achieves state-of-the-art performance (outperforms Tacotron2 with a gap of 0.048) and is very close to human quality (4.39 vs 4.44 in MOS).
Background and objective:
Due to the development in digital microscopic imaging, image processing and classification has become an interesting area for diagnostic research. Various techniques are available in the literature for the detection of Acute Lymphocytic Leukemia from the single cell blood smear images. The purpose of this work is to develop an effective method for leukemia detection.
Methods:
This work has developed deep learning based leukemia detection module from the blood smear images. Here, the detection scheme carries out pre-processing, segmentation, feature extraction and classification. The segmentation is done by the proposed Mutual Information (MI) based hybrid model, which combines the segmentation results of the active contour model and fuzzy C means algorithm. Then, from the segmented images, the statistical and the Local Directional Pattern (LDP) features are extracted and provided to the proposed Chronological Sine Cosine Algorithm (SCA) based Deep CNN classifier for the classification.
Results:
For the experimentation, the blood smear images are considered from the AA-IDB2 database and evaluated based on metrics, such as True Positive Rate (TPR), True Negative Rate (TNR), and accuracy. Simulation results reveal that the proposed Chronological SCA based Deep CNN classifier has the accuracy of 98.7%.
Conclusions:
The performance of the proposed Chronological SCA-based Deep CNN classifier is compared with the state-of-the-art methods. The analysis shows that the proposed classifier has comparatively improved performance and determines the leukemia from the blood smear images.
The generalization error bound of the support vector machine (SVM) depends on the ratio of the radius and margin. However, conventional SVM only considers the maximization of the margin but ignores the minimization of the radius, which restricts its performance when applied to joint learning of feature transformation and the SVM classifier. Although several approaches have been proposed to integrate the radius and margin information, most of them either require the form of the transformation matrix to be diagonal, or are nonconvex and computationally expensive. In this paper, we suggest a novel approximation for the radius of the minimum enclosing ball in feature space, and then propose a convex radius-margin-based SVM model for joint learning of feature transformation and the SVM classifier, i.e., F-SVM. A generalized block coordinate descent method is adopted to solve the F-SVM model, where the feature transformation is updated via the gradient descent and the classifier is updated by employing the existing SVM solver. By incorporating with kernel principal component analysis, F-SVM is further extended for joint learning of nonlinear transformation and the classifier. F-SVM can also be incorporated with deep convolutional networks to improve image classification performance. Experiments on the UCI, LFW, MNIST, CIFAR-10, CIFAR-100, and Caltech101 data sets demonstrate the effectiveness of F-SVM.
As a machine learning algorithm, AdaBoost has obtained considerable success in data classification and object detection. Later its generalized version called Real AdaBoost was proposed by Schapire and Singer. Real AdaBoost increases weights for misclassified samples and decreases weights for correctly classified samples in every iteration. This kind of weight adjustment focuses on the samples with large weights and tries to make them correctly classified in future runs. However, it may lead to the misclassification of some other samples that have been correctly classified in previous runs. If we can curb this kind of misclassification during the boosting process, a faster training can be achieved. Based on this assumption, we propose Parameterized AdaBoost in which a parameter is devised to penalize the misclassification of samples that have already been correctly classified. Then we analyse that samples with positive classification margins in Parameterized AdaBoost are more than in Real AdaBoost. Experimental results show that our approach achieves a faster convergence of the training error and also improves the generalization error to some degree when compared with Real AdaBoost.
This communication introduces a hybrid invasive weed optimization and particle swarm optimization (IWO/PSO) algorithm dedicated to pattern synthesis of conformal arrays. A 3 × 9 cylindrical conformal microstrip array is studied to demonstrate the proposed algorithm. The active element pattern of each element in the conformal array is extracted and the phase and amplitude of each excited voltage are optimized to achieve the required goals using the hybrid IWO/PSO algorithm. The results indicate that, compared with standard IWO and PSO algorithms, the introduced hybrid algorithm maintains the respective advantages of the IWO and PSO algorithms while excluding their respective deficiencies. Moreover, the hybrid algorithm is very reliable in achieving global optimality.