Content uploaded by Dr. Uday Pratap Singh
Author content
All content in this area was uploaded by Dr. Uday Pratap Singh on Feb 11, 2018
Content may be subject to copyright.
Image Segmentation Using Computational Intelligence Techniques: Review
Abstract
Image segmentation methodology is a part of nearly all computer schemes as a pre-processing phase to excerpt more meaningful and useful information for analysing the objects within an image. Segmentation of an image is one of the most conjoint scientific matter, essential technology and critical constraint for image investigation and dispensation. There has been a lot of research work conceded in several emerging algorithms and approaches for segmentation, but even at present, no solitary standard technique has been proposed. The methodologies present are broadly classified among two classes i.e. traditional approaches and Soft computing approaches or Computational Intelligence (CI) approaches. In this article, our emphasis is to focus on Soft Computing (SC) techniques which has been adopted for segmenting an image. Nowadays, it is quite often seen that SC or CI is cast-off frequently in Information Technology and Computer Technology. However, Soft Computing approaches working synergistically provides in anyway, malleable information processing competence to manipulate real-life enigmatic circumstances. The impetus of these methodologies is to feat the lenience for ambiguity, roughness, imprecise acumen and partial veracity for the sake to attain compliance, sturdiness and economical results. Neural Networks (NNs), Fuzzy Logic (FL), and Genetic Algorithm (GA) are the fundamental approaches of SC regulation. SC approaches has been broadly implemented and studied in the number of applications including scientific analysis, medical, engineering, management, humanities etc. The paper focuses on introducing the various SC methodologies and presenting numerous applications in image segmentation. The acumen is to corroborate the probabilities of smearing computational intelligence to segmentation of an image. The available articles about usage of SC in segmentation are investigated, especially focusing on the core approaches like FL, NN and GA and efforts has been also made for collaborating new techniques like Fuzzy C-Means from FL family and Deep Neural Network or Convolutional Neural Network from NN family. The impression behind this work is to simulate core Soft Computing methodologies, along with encapsulating various terminologies like evaluation parameters, tools, databases, noises etc. which can be advantageous for researchers. This study also identifies approaches of SC being used, often collectively to resolve the distinctive dilemma of image segmentation, concluding with a general discussion about methodologies, applications followed by proposed work.
... There is a lack of high-quality segmentation datasets, and existing datasets often suffer from limited scene diversity [28,29]. Additionally, segmentation techniques require significant computational resources [33], which may lead to increased costs [34]. ...
Effective safety management is vital for ensuring construction safety. Traditional safety inspections in construction heavily rely on manual labor, which is both time-consuming and labor-intensive. Extensive research has been conducted integrating computer-vision technologies to facilitate intelligent surveillance and improve safety measures. However, existing research predominantly focuses on singular tasks, while construction environments necessitate comprehensive analysis. This study introduces a multi-task computer vision technology approach for the enhanced monitoring of construction safety. The process begins with the collection and processing of multi-source video surveillance data. Subsequently, YOLOv8, a deep learning-based computer vision model, is adapted to meet specific task requirements by modifying the head component of the framework. This adaptation enables efficient detection and segmentation of construction elements, as well as the estimation of person and machine poses. Moreover, a tracking algorithm integrates these capabilities to continuously monitor detected elements, thereby facilitating the proactive identification of unsafe practices on construction sites. This paper also presents a novel Integrated Excavator Pose (IEP) dataset designed to address the common challenges associated with different single datasets, thereby ensuring accurate detection and robust application in practical scenarios.
Diseases and pests in plants/crops are major causes of significant agricultural losses with economic, social and ecological impacts. Therefore, there is a need for early identification of plant diseases and pests through automated systems. Recently, machine learning-based methods have become popular in solving agricultural problems such as plant diseases faced by technically-noob farmers. This work proposes a novel method based on stacking ensemble machine learning to detect plant diseases in Uradbean precisely. Two classifiers: support vector machine (SVM), random forest (RF) are trained on a dataset consists of Uradbean infected and healthy leaf images. These classifiers are stacked with logistic regression (LR) classifier. In the diverse ensemble, LR classifier is used as a meta-learner which enhanced the precision of the disease classification. The fuzzy C-Means clustering with particle swarm optimization is used for image segmentation. Haralick, Hu Moments and color histogram methods are used in feature extraction. During the tests, the proposed model is also compared with pre-trained networks: DenseNet-201, ResNet-50, and VGG19. It achieved an impressive classification accuracy of 96.82 % which is higher than the individual classifiers and pre-trained networks. To validate model performance, it is evaluated on a benchmark public dataset consists of Apple leaf images and achieved 98.30% accuracy. It is observed that ensemble method reflects an advantage over individual models in increasing the classification rates and reducing the computational overhead in comparison to pre-trained networks which struggle due to the issues such as irrelevant features, generation of pertinent characteristics, and noise
Brain tumor diagnosis has been revolutionized with the advent of deep learning technique: CNN. The chapter explores the application of CNN in the medical diagnosis of brain tumor using MRI and CT scan images. Initially, the simplified explanation of CNN with basic architecture is shown. Later, the operational mechanism of CNN is explained which is which serves in brain tumor detection with high accuracy and precision. It mimics human perception and analyzes intricate details within images that signify the presence of an ailment. In the later part, the concept of brain tumors is discussed along with the importance of early detection of brain tumors is also highlighted outlining the impact on individuals. In the subsequent part, the training process of CNN to detect brain tumors is discussed to equip readers with the requisite knowledge and skills to train the model. Demonstrating the relationship between CNN and medical imaging techniques, this chapter aims to reduce the complexity in the process of brain tumor detection, highlighting the transformative potential of CNN in healthcare services.
Plant disease identification plays a crucial role in agricultural management, but the traditional methods are time-consuming and imprecise. This work contributes to the ongoing efforts to develop reliable and efficient solutions for automated plant disease diagnosis, ultimately aiding in the timely management and mitigation of agricultural challenges. This study investigated the effectiveness of various deep convolutional neural networks (CNNs) for automated plant disease identification from leaf images. By exploring deep and transfer learning techniques such as CNNs, InceptionNet, DenseNet 121, and ResNet-50, a Deep Convolutional Neural Network (DCNN) is proposed to categorize the leaf disease. Different activation functions, specifically Rectified Linear Unit (ReLU) and Swish, are employed to investigate their impact on model performance. The experiments revealed that the DCNN architecture, when paired with the Swish activation function, demonstrated 96% accuracy in plant disease identification.
This chapter introduces the different aspects of computer vision and drone technology in the field of agriculture. The idea is to present various concepts, theories, applications, and future advancement of these underlying technologies. We hope that this chapter will help the readers to get the basic insights and their practical adaptability in agriculture. Further, to start with, the various terms of computer vision methodologies are discussed first, and then in the later part, the core of drone technology is introduced.
A novel method of Lung image segmentation based on Textural Echo State Neural Network is proposed in this paper. This work combines the textural features of different pixels in the lung images in order to increase the performance of region of interest segmentation. For segmenting lung images, a two-step process is done. First, an unsupervised border detection is done in order to separate the foreground lung images from the background. This process begins by identifying the textural features that represent the lung border clearly by the process of Texture analysis. In the second step, the obtained features are further given as input towards the Echo state neural networks as training pattern in order to obtain the segmented lung region. The proposed algorithm is trained and tested for 450 skin lesion images in order to evaluate its accuracy. An average accuracy of 97.8% for segmenting lung images has been obtained.
Automatic skin lesion segmentation in dermoscopic images is a challenging task due to the low contrast between lesion and the surrounding skin, the irregular and fuzzy lesion borders, the existence of various artifacts, and various imaging acquisition conditions. In this article, we present a fully automatic method for skin lesion segmentation by leveraging a 19-layer deep convolutional neural networks (CNNs) that is trained end-to- end and does not rely on prior knowledge of the data. We propose a set of strategies to ensure effective and efficient learning with limited training data. Furthermore, we design a novel loss function based on Jaccard distance to eliminate the need of sample re-weighting, a typical procedure when using cross entropy as the loss function for image segmentation due to the strong imbalance between the number of foreground and background pixels. We evaluated the effectiveness, efficiency, as well as the generalization capability of the proposed framework on two publicly available databases. One is from ISBI 2016 Skin Lesion Analysis Towards Melanoma Detection Challenge, and the other is the PH2 database. Experimental results showed that the proposed method outperformed other state-of-the-art algorithms on these two databases. Our method is general enough and only needs minimum pre- and post-processing, which allows its adoption in a variety of medical image segmentation tasks.
Medical image processing plays an important role in supporting the diagnosis of various diseases. Brain magnetic resonance imaging (MRI) image is widely used to support the decisions from doctors who will decide if there are any issues in a brain. The essence of the MRI is segmentation which is the basic for damaged area selection, quantitative measurement and 3-dimensional reconstruction. In order to effectively identify the located objects, this paper introduces a segmentation algorithm using global entropy minimization. This algorithm uses two times segmentation approach based on the cluster area image model to overcome the negative influences of shifted segmentation. From the experiments, the proposed algorithm get the best performance and keeps the highest accuracy. For the similarity, the proposed algorithm has almost the same performance of least biased fuzzy clustering (LBFC) which have 10% out performance on fuzzy C-means algorithm (FCMA).
Conventional traffic light detection methods often suffers from false positives in urban environment because of the complex backgrounds. To overcome such limitation, this paper proposes a method that combines a conventional approach, which is fast but weak to false positives, and a DNN, which is not suitable for detecting small objects but a very powerful classifier. Experiments on real data showed promising results.
Magnetic Resonance Imaging is the preferred imaging modality for assessing brain tumors, and segmentation is necessary for diagnosis and treatment planning. Thus, robust automatic segmentation methods are required. Machine learning proposals where the model is learned from data are quite successful. Hierarchical segmentation approaches firstly segment the whole tumor, followed by intra-tumor tissue identification. However, results comparing it with single stages approaches are needed, as state of the art results are also achieved by all-at-once strategies. Currently, fully convolutional networks approaches for segmentation are very efficient. In this paper, a hierarchical approach for brain tumor segmentation using a fully convolutional network is studied. The evaluation is performed on the Brain Tumor Segmentation Challenge 2013 dataset, and we report the metrics Dice Score Coefficient, Positive Predictive Value, and Sensitivity. Results show benefits from segmenting the complete tumor first, over all tissues in one stage. Moreover, the tumor core also benefits from such approach. This behavior may be justified by the high data imbalance observed between tumor and normal tissues, which is mitigated by considering the tumor as a whole.
Remote Sensing is a technique to obtain data from objects without physical contact using sensors. If it is performed by hyperspectral sensors, then is possible to have a large number of bands that allows many applications. In this case, a large amount of data needs to be processed and analyzed, and therefore the practical use of these images have challenges with storage, transport and processing time. Another issue is that many of these bands contain redundant information due to high correlation between them that can cause instability in the convergence of computing processes. Because of these problems there is a need to select important bands to be processed. In many monitoring applications through remote sensing prior knowledge of areas can be used to improve the processing. Algorithms that use this information are known as supervised methods. This study introduces a methodology for band selection based on Genetic Algorithms and Gaussian Maximum Likelihood Classifier. The results were significant and demonstrate that the proposed method can reduce about 26,56% of the number of bands in hyperspectral images maintaining good quality for interpretation. The result of the similarity of the segmentation of images generated with fewer bands and more bands in the proposed methodology and the other algorithm were also compared. With the results expected to apply the proposed methodology in practical work and further study.