Proposed Segmentation Technique Research Articles

Lesion detection and classification in dermoscopic images using optimal threshold based on Newton Raphson iterative method

Automatic image thresholding is commonly used in the area of computer vision and pattern recognition for object detection. This research work proposes a new method of lesion detection in dermoscopic images using an optimal threshold based on the Newton-Raphson iteration algorithm for diagnosing melanoma. The proposed method incorporates a new strategy of finding the optimal threshold using first-order and second-order edge derivatives. The first- and second-order derivatives values obtained from the images are applied in the Newton-Raphson method which, in turn, converges to a value after a number of iterations. The converged value is considered the optimal threshold and used as a parameter for the separation of lesion regions from healthy skin in dermoscopic images. To test the proposed system performance, the experiment is carried out on two standard data sets, the ISIC Archive and the PH2, of dermoscopic images using different classifiers such as the Naive Bayes (NB), AdaBoost and Bag of Visual words (BOVW). The proposed segmentation technique segments the lesion regions from skin images and Bag of visual words classifier classifies the dermoscopic images into different classes namely common nevi, atypical nevi and melanoma. The proposed thresholding-based segmentation performance is compared against the Otsu method and Otsu using genetic algorithm. The experimental findings reveal that the proposed Newton-Raphson iterative method surpasses various state-of-the-art and recent methods in relation to computational speed, error rate, sensitivity, and accuracy.

Aquila-particle swarm based cooperative search optimizer with superpixel techniques for epithelial layer segmentation

The segmentation of epithelial layers from oral histopathology images plays a crucial role for early detection of oral cancer disease. As a result, more accurate segmentation of this layer is of the utmost importance for a Computer Aided Diagnosis (CAD) system. Therefore, this paper presents a superpixel image-based clustering technique using an improved Nature-Inspired Optimization Algorithm (NIOA), called the Cooperative Search (CS) algorithm. Here, superpixel image is utilized to construct a faster and noise-resistant oral histopathology image clustering method. The designed CS, on the other hand, is based on two popular NIOAs, the Aquila Optimizer (AO) and the Particle Swarm Optimizer (PSO), has better exploration-exploitation abilities. This CS has been used in the image clustering field to circumvent the problem of local optima trapping. On the other hand, a comparative investigation of three popular superpixel strategies with the proposed CS optimizer has been conducted to determine the optimal superpixel strategy for epithelial layer segmentation from clean as well as noisy oral histopathology images. Finally, the optimal superpixel strategy with CS has been tested with other state-of-the-art image segmentation techniques in the epithelial layer segmentation domain. The results obtained by the proposed segmentation technique are 98.82%, 96.70%, 97.60%, and 95.25%, in terms of average Accuracy, MCC, Dice, and Jaccard, respectively, which all are better than the cutting-edge segmentation methods. The proposed segmentation model is also evaluated over the leaf segmentation model and achieved better results visually and numerically. The optimization efficiency of the CS has also been measured over CEC2019 mathematical benchmark test functions and produced competitive results compared to other tested NIOAs.

A Hybrid Method for Magnetic Resonance Brain Images Classification and Segmentation Using Soft Computing Techniques

Nowadays, Brain tumor is a serious life-threatening disease that can often be treated with risky surgeries. Various classification and segmentation methods for MR (Magnetic Resonance) brain images have been proposed but the expected accuracy value could not be reached so far. In this paper, we proposed a hybrid approach that includes modified fuzzy C-means and ANN classifier. It consists of five stages (a) Noise removal (b) Feature extraction (c) Feature selection (d) Classification (e) Segmentation. Initially, a genetic optimized median filter (GOMF) is used to remove noise present in the input image, and then the essential features are extracted and selected using Discrete Wavelet Transform (DWT) & Principle Component Analysis (PCA) algorithms respectively. The normal and abnormal images are classified using the ANN classifier. Finally, it is processed through a Modified fuzzy C-means algorithm to segment the tumor portion separately. The proposed segmentation technique has been tested on the BRATS dataset and produces a sensitivity of 98%, Jaccard index of 97%, specificity of 98%, and accuracy of 95%.

A novel technique for dental radiographic image segmentation based on neutrosophic logic

Oral diseases are very prevalent worldwide and affect people of all ages. Dentists depend on x-rays to study the characteristics of oral diseases. The segmentation and analysis of dental X-ray images pose various challenges compared to other medical images. This makes dental X-ray imaging more challenging because of poor resolution, which makes the segmentation of different parts of teeth and their abnormalities unreliable. It has been shown that Dental X-ray Image Segmentation (DXIS) is a primary and critical step in obtaining relevant and important details about oral diseases. DXIS plays a crucial role in practical dentistry to help identify various periodontal diseases. The proposed methodology automatically segments the teeth regions and assists in further analysis. It works on both peri-apical and panoramic types of dental radiographic images. Neutrosophic logic is used to select the initial region of interest. The best way to improve the performance and make the system faster is to restrict the computation within the foreground regions. The input dental radiographic image is mapped into the neutrosophic domain using the patch level feature, gradient feature, entropy feature, and local binary pattern. Applying neutrosophic logic helps to localize the initial region of interest. Subsequently, a fuzzy c-means algorithm is applied to segment a more accurate region of interest. The proposed methodology has been evaluated on publicly available data sets, ‘Panoramic Dental X-rays with Segmented Mandibles’ and ‘Digital Dental X-ray Database for Caries Screening,’ with the result that the accuracy of the proposed methodology is as high as 93.20%. This performance level confirms that the proposed segmentation technique highly correlates with the manual system.

Outlier Based Skimpy Regularization Fuzzy Clustering Algorithm for Diabetic Retinopathy Image Segmentation

Blood vessels are harmed in diabetic retinopathy (DR), a condition that impairs vision. Using modern healthcare research and technology, artificial intelligence and processing units are used to aid in the diagnosis of this syndrome and the study of diagnostic procedures. The correct assessment of DR severity requires the segmentation of lesions from fundus pictures. The manual grading method becomes highly difficult and time-consuming due to the wide range of the morphologies, number, and sizes of lesions. For image segmentation, traditional fuzzy clustering techniques have two major drawbacks. First, fuzzy memberships based clustering are more susceptible to outliers. Second, because of the lack of local spatial information, these techniques often result in oversegmentation of images. In order to address these issues, this research study proposes an outlier-based skimpy regularization fuzzy clustering technique (OSR-FCA) for image segmentation. Clustering methods that use fuzzy membership with sparseness can be improved by incorporating a Gaussian metric regularisation into the objective function. The proposed study used the symmetry information contained in the image data to conduct the image segmentation using the fuzzy clustering technique while avoiding over segmenting relevant data. This resulted in a reduced proportion of noisy data and better clustering results. The classification was carried out by a deep learning technique called convolutional neural network (CNN). Two publicly available datasets were used for the validation process by using different metrics. The experimental results showed that the proposed segmentation technique achieved 97.16% and classification technique achieved 97.26% of accuracy on the MESSIDOR dataset.

Human Coronary Plaque Optical Coherence Tomography Image Repairing, Multilayer Segmentation and Impact on Plaque Stress/Strain Calculations.

Coronary vessel layer structure may have a considerable impact on plaque stress/strain calculations. Most current plaque models use single-layer vessel structures due to the lack of available multilayer segmentation techniques. In this paper, an automatic multilayer segmentation and repair method was developed to segment coronary optical coherence tomography (OCT) images to obtain multilayer vessel geometries for biomechanical model construction. Intravascular OCT data were acquired from six patients (one male; mean age: 70.0) using a protocol approved by the local institutional review board with informed consent obtained. A total of 436 OCT slices were selected in this study. Manually segmented data were used as the gold standard for method development and validation. The edge detection method and cubic spline surface fitting were applied to detect and repair the internal elastic membrane (IEM), external elastic membrane (EEM) and adventitia-periadventitia interface (ADV). The mean errors of automatic contours compared to manually segmented contours were 1.40%, 4.34% and 6.97%, respectively. The single-layer mean plaque stress value from lumen was 117.91 kPa, 10.79% lower than that from three-layer models (132.33 kPa). On the adventitia, the single-layer mean plaque stress value was 50.46 kPa, 156.28% higher than that from three-layer models (19.74 kPa). The proposed segmentation technique may have wide applications in vulnerable plaque research.

An Instrumented Glove for Recognition of Brazilian Sign Language Alphabet

This paper presents the development and analysis of a system for Brazilian Sign Language (Libras) recognition, which is composed of an instrumented glove and an acquisition system. The instrumented glove has five flex sensors, an inertial sensor, and two contact sensors (made with silver-coated fabric). For data acquisition, five volunteers performed gestures in Libras. In the proposed segmentation technique, the collected data were segmented into three periods (fixed amounts of signal samples): construction period, alphabet gesture, and relaxation period. The periods of alphabet gesture are submitted to segmentation, and the classifier was designed to recognize the 26 letters of alphabet. Different classifiers were employed to classify the gestures and analyze the system in different scenarios. For the classification in group of sensors, the highest performance obtained was an accuracy of 96.15% with Random Forest (RF) classifier.

Morphological Based Medical Image Processing on Cervical Cytology Cancer Images Using Connected Component Techniques.

Cervical cancer is the fourth most common cancer in women. In 2018, it was estimated 570000 women were diagnosed with cervical cancer worldwide and about 311000 women died from the disease. An efficient technique is essential for solving the complication in the diagnosis of cervical cancer images. In this research, a new method is developed for cervical cancer image segmentation. First, the RGB image is converted into an HSI color model. The thresholding is applied to the saturation and intensity components to get binary images. These binary images are combined to get a new mask. Using the connected component concept nucleus and cytoplasm are segmented accurately. For the performance evaluation, peak signal-to-noise ratio (PSNR), mean square error (MSE), structural similarity index (SSIM), image quality index (IQI), structural content (SC), normalized cross-correlation (NK), precision (PR), recall (RC), the Average Difference (AD) and image fidelity (IF) are taken. The proposed techniques' highest PSNR values are 44.2341, 46.7953, 60.5925, and 61.4862, respectively. The proposed segmentation technique can attain a high PSNR (>40db) value at a threshold value equal to 0.1. Also proposed approach attains good precision, recall, and SSIM values. The lowest MSE values using proposed segmentation techniques are 0.0454, 0.0351, 0.0924, and 0.0271 individually. The AD, NK, SC, NAE, and LMSE values for the implemented approach are low, showing the segmented image's quality is very good. Thus, the proposed model performed better compared to other methods.

Adaptive Image Segmentation for Traumatic Brain Haemorrhage

It is challenging to establish a significant solution with computer techniques to improve the speed and efficiency of Traumatic Brain Injury (TBI) diagnosis. Several segmentation methods involving diverse precision and a degree of effort have been proposed and detailed within the related literature. Segmentation of Brain image is one of the significant clinical diagnostics implements. This paper proposes a modified (MDRLSE) calculation for haemorrhage segmentation on Computed Tomography (CT) images. The image noise that abdicates the obscured edges is utilized to portray the precise boundary of the haemorrhage region. The proposed segmentation technique achieved an accuracy rate of 97.16%. The technique is implemented using an edge-based involved contour model for image segmentation, providing a simple narrowband to significantly reduce computational costs. The performance results show that it is effective for TBI image segmentation in brain images with various characteristics.

Adaptive Image Segmentation for Traumatic Brain Haemorrhage

It is challenging to establish a significant solution with computer techniques to improve the speed and efficiency of Traumatic Brain Injury (TBI) diagnosis. Several segmentation methods involving diverse precision and a degree of effort have been proposed and detailed within the related literature. Segmentation of Brain image is one of the significant clinical diagnostics implements. This paper proposes a modified (MDRLSE) calculation for haemorrhage segmentation on Computed Tomography (CT) images. The image noise that abdicates the obscured edges is utilized to portray the precise boundary of the haemorrhage region. The proposed segmentation technique achieved an accuracy rate of 97.16%. The technique is implemented using an edge-based involved contour model for image segmentation, providing a simple narrowband to significantly reduce computational costs. The performance results show that it is effective for TBI image segmentation in brain images with various characteristics.

MRI Image Segmentation Using Bat Optimization Algorithm with Fuzzy C Means (BOA-FCM) Clustering

Functional and anatomical information extraction from Magnetic Resonance Images (MRI) is important in medical image applications. The information extraction is highly influenced by the artifacts in the MRI images. The feature extraction involves the segmentation of MRI images. We present a MRI image segmentation using Bat Optimization Algorithm (BOA) with Fuzzy C Means (FCM) clustering. Echolocation of bats is utilized in Bat Optimization Algorithm. The proposed segmentation technique is evaluated with existing segmentation techniques. Results of experimentation shows that proposed segmentation technique outperforms existing methods and produces 98.5% better results.

Dynamically learned PSO based neighborhood influenced fuzzy c-means for pre-treatment and post-treatment organ segmentation from CT images

Background and ObjectiveThe accurate segmentation of pre-treatment and post-treatment organs is always perceived as a challenging task in medical image analysis field. Especially, in those situations where the amount of data set is limited, the researchers are compelled to design unsupervised model for segmentation. In this paper, we propose a novel dynamically learned particle swarm optimization based neighborhood influenced fuzzy c-means (DLPSO-NIFCM) clustering (unsupervised learning model) for solving pre-treatment and post-treatment organs segmentation problems. The proposed segmentation technique has been successfully applied to segment the liver parts from the Computed Tomography (CT) images of abdomen and also the lung parenchyma from the lungs CT images. MethodologyIn the proposed method, we formulate a primary convex objective function by considering the membership value of a pixel as well as the membership of its other neighboring pixels. Then we apply a new algebraic transformation on the primary objective function to design a new and more suitable objective function without losing convexity of the primary objective function. This new objective function is compatible for hybridization with any heuristic search technique in true sense. In this work, we propose a dynamically learned PSO to obtain the initial cluster centroids from the final objective function. Finally, we use a graph-based isolation mechanism for refining the segmentation results. Results and ConclusionThis hybrid method, along with the restructured single variable objective function of the distance, leads to accurate clustering results with relatively lesser converging time as compared to the state-of-the-art methods. The segmentation results, obtained through several experiments with real CT images, are encouraging. The numerical values of different performance metrics obtained over the same data set confirm that the proposed algorithm performs better with respect to the state-of-the-art methods. Hence, we may consider the proposed method as a promising tool for clustering and CT image segmentation in a Computer Aided Diagnostic (CAD) system.

Automated grading of mangoes based on surface defect detection using a combined approach of image segmentation

Image segmentation is an essential and critical step in huge number of applications of image processing. Accuracy of image segmentation influence retrieved information for further processing in classification and other task. In image segmentation algorithms, a single segmentation technique is not sufficient in providing accurate segmentation results in many cases. In this paper we are proposing a combining approach of image segmentation techniques for improving segmentation accuracy. As a case study fruit mango is selected for classification based on surface defect. This classification method consists of three steps: (a) image pre-processing, (b) feature extraction and feature selection and (c) classification of mango. Feature extraction phase is performed on an enhanced input image. In feature selection PCA methodology is used. In classification three classifiers BPNN, Naïve bayes and LDA are used. Proposed image segmentation technique is tested on online dataset and our own collected images database. Proposed segmentation technique performance is compared with existing segmentation techniques. Classification results of BPNN in training and testing phase are acceptable for proposed segmentation technique.

Intuitionistic Level Set Segmentation for Medical Image Segmentation

Introduction: Image segmentation is one of the basic practices that involve dividing an image into mutually exclusive partitions. Learning how to partition an image into different segments is considered as one of the most critical and crucial step in the area of medical image analysis. Objective: The primary objective of the work is to design an integrated approach for automating the process of level set segmentation for medical image segmentation. This method will help to overcome the problem of manual initialization of parameters. Methods: In the proposed method, input image is simplified by the process of intuitionistic fuzzification of an image. Further segmentation is done by intuitionistic based clustering technique incorporated with local spatial information (S-IFCM). The controlling parameters of level set method are automated by S-IFCM, for defining anatomical boundaries. Results: Experimental results were carried out on MRI and CT-scan images of brain and liver. The results are compared with existing Fuzzy Level set segmentation; Spatial Fuzzy Level set segmentation using MSE, PSNR and Segmentation Accuracy. Qualitatively results achieved after proposed segmentation technique shows more clear definition of boundaries. The attain PSNR and MSE value of propose algorithm proves the robustness of algorithm. Segmentation accuracy is calculated for the segmentation results of the T-1 weighted axial slice of MRI image with 0.909 value. Conclusion: The proposed method shows good accuracy for the segmentation of medical images. This method is a good substitute for the segmentation of different clinical images with different modalities and proves to give better result than fuzzy technique.

Tumor Segmentation by a Self-Organizing-Map based Active Contour Model (SOMACM) from the Brain MRIs

Segmentation of tumors from the brain Magnetic Resonance Images (MRIs) is very important for the analysis and right treatment. Tumors treated at early stages improve the survival time. This paper proposes an advanced method named SOMACM which is a combination of Self -Organizing -Map (SOM) and Active Contour Model (ACM) for the efficient segmentation of brain MRIs to detect tumors. ACM is an energy-based segmentation method and treats the segmentation as an optimization issue. It can model complex shapes and handles topological changes in the object boundary. The customary ACMs rely upon the intensities of the pixels and are very vulnerable to parameter tuning hence it is very difficult to segment the images of distinct pixel intensities. ACMs will evolve from the object boundary for the images consisting of Intensity Inhomogeneity (IIH). Neural Networks (NNs) are exceptionally compelling in processing the images of inhomogeneities. Furthermore, image segmentation can be done by NNs without the use of an objective function. The proposed SOMACM method works by precisely incorporating the global information extracted from the weights of the trained SOM neurons which helps in modeling complex shapes and distinct intensity distributions. It can handle images with noise, intensity similarity and IIH. The proposed segmentation technique is not sensitive to parameter tuning. The outcomes of the proposed SOMACM demonstrate the improved accuracy in the segmentation results of different types of tumor images, in contrast with the individual SOM, ACM, Fuzzy-C- Means (FCM), Particle Swarm Optimization (PSO) and Probabilistic Neural Networks (PNN) segmentation methods.

Design of a histopathology image segmentation algorithm for CAD of cancer

In the present world, medical diagnosis systems depend a lot on medical imaging and digital nosology. The digital nosology is leading to computer-aided diagnosis to facilitate the rapid and accurate screening of specimens. In this work, we present a method based on object shaped kernel for histopathology image segmentation to identify and analyze the cell nuclei region. The proposed segmentation technique is a multi-stage procedure, which includes contrast enhancement of the image, nuclei region extraction, nucleus centroid marking, nuclei area refinement and, complex nuclei separation. The performance of proposed method is measured using three standard H&E stained histopathology image datasets and a proposed dataset. The simulation results of the segmentation schemes are evaluated using the F1-score, Hausdorff distance and Jaccard index. The proposed system also introduced a novel performance measure as cumulative factor. The superiority of proposed segmentation system is verified in terms of cumulative factor in comparison to various the state of art methods.

Computer-aided system for Leukocyte nucleus segmentation and Leukocyte classification based on nucleus characteristics

Differential count of leukocytes plays a consequential role in the determination of diagnostic details of a patient. Conventionally, leukocyte identification is performed manually by skilled medical personnel. But the involvement of humans slows down the task and also affects the process outcome. Designing a computer-aided system to automate this task can reduce the load on the medical personnel and help in attaining precise results in minimal time. This research presents an image-processing technique to segment the Leukocyte Nucleus from the blood smear image and categorize them into their constituent classes based on the features drawn from the segmented region. Mathematical Operations are performed on the color channels of RGB and LAB color spaces to obtain Nucleus-enhanced grayscale Image. Leukocyte Nucleus is segmented using Fuzzy C-Means Clustering. Geometrical, Color and Texture features are extracted from the segmented nucleus. Feed-Forward Neural Network and Support Vector Machine classifiers are employed to classify the leukocytes into its constituent classes. The proposed segmentation technique rendered better segmentation performance compared to the methods reviewed in the literature with an overall segmentation accuracy of 88.1%. The Feed Forward Neural Network classifier yielded an accuracy of 92.8% for Leukocyte classification.

A feature extraction based trajectory segmentation approach based on multiple movement parameters

Analyzing the trajectories of movements among moving objects is of interest in many fields to understand the dynamics and behavior of those objects. In this paper, a cluster-centric trajectory segmentation approach is proposed to reveal and visualize segments of trajectories among moving objects. Characteristics such as position, direction, and speed of moving objects (called movement parameters) are considered for this purpose. First, profiles generated for different movement parameters are divided into several portions using sliding windows of different length. Next, changes with respect to each particular movement parameter profile in the sliding windows are extracted as features. Finally, by clustering the extracted features, subsequences of trajectories with similar movement characteristics are detected. Some cluster-validity indices were used to find the (near) optimal number of clusters. The performance of the proposed segmentation technique is evaluated through a trajectory clustering as well as a movement pattern detection case study over some real-word datasets.

A fast and automated segmentation method for detection of masses using folded kernel based fuzzy c-means clustering algorithm

Present study proposes a fast, accurate and automated segmentation approach of mammographic images using kernel based fuzzy c-means (FCM) clustering technique. This approach exploits the significant regional features of mammograms which address the properties of different breast densities. The proposed segmentation approach captures those regional features using appropriate kernel and hence apply fuzzy clustering technique for segmenting the masses. This study also introduces kernel based FCM (KFCM) approach in a folded way to process a combination of significant features simultaneously. Suitable choice of kernel size also assists to collect all possible variations of regional features with minimum blocking effect in the output results. Performances of the proposed methodology are analyzed qualitatively and quantitatively in compare to other clustering-based segmentation techniques. Since the proposed approach is able to resolve uncertain and imprecise characteristics of mammograms, it performs superior to other techniques. Convergence time of the proposed method is also assessed and compared with other conventional clustering techniques. Kernel based approach of the proposed segmentation technique reduces the number of data points for clustering and hence convergence speed improves over the conventional algorithms. This study also shows a variation of convergence speed of the proposed segmentation method with different image sizes.

Brain Tumor Analysis in Basf Framework

The medical image processing is extensively used in several areas. In earlier detection and treatment of these diseases is very helpful to find out the abnormality issues in that image. Here there are number of methods available for diagnosis to detect the brain tumor of MRI image. This study deals with there are two main contributions are implemented in this filter method. (1)The extension of adaptive bilateral method to apply sub-bands of low frequency signal decomposed using wavelet transform. A wavelet threshold is combined with adaptive bilateral method to form an innovative structure in image de-noising method. It’s very efficient to eliminate noise in original noisy images. (2) First detected block boundary and texture regions discontinuities to adapt or control the parameters of spatial and intensity in bilateral filter. The adaptive method can improve the restored image quality in this test result compared with standard bilateral filter. The proposed segmentation technique uses novel strip method and the image is split into number of strips 3, 4, 5 and 6. Using a hybrid Assured Convergence PSO (ACPSO) and Fuzzy C-Mean Clustering (FCM) was proposed method. The segmentation algorithm presented in this research gives 95% of accuracy rate to detect brain tumor when strip count is 5. In this research work presented a feature vector using a different statistical texture analysis of brain tumor from MRI image. The statistical feature texture is computed using GLCM (Gray Level Co-occurrence Matrices) of Brain Nodule structure. For this research work, the brain nodule segmented using strips method to implemented marker watershed image segmentation based on PSO (Particle Swarm Optimization) and Fuzzy C-means clustering (FCM). Furthermore, the four angles 0o, 45o, 90o and 135o are calculated the segmented brain image in GLCM. The four angular directions are calculated using texture features are correlation, energy, contrast and homogeneity. The accuracy rate of previous method was compared and proved the proposed method an Assured Convergence Particle Swarm Optimization (ACPSO)-Fuzzy C-Mean (FCM) and using SVM classification technique is suitable for the early detection of brain tumor. In proposed, a tumor extraction is improved in ASPSO-FCM and SVM classification with better accuracy rate of 95.31%.