MIT-World Peace University (MIT-WPU) is a top educational institute located in Pune, India. It is recognized within&nbsp;UGC under Government of Maharashtra Act No. XXXV 2017.&nbsp;

It is part of the MIT group of institutions (established in 1983) which covers 10+ campuses across India and has more than 50,000+ students enrolled across various disciplines at a given point of time. The institute houses 15 schools ranging from Engineering &amp; Technology, Design, Pharmacy, Commerce, Governance among others. It offers undergraduate, postgraduate and PhD programs among different disciplines.&nbsp;

&nbsp;

MIT World Peace University

Texture segmentation is one of the popular researchdomains and researchers across the globe are working on texturesegmentation to enhance segmentation performance to address itsrequirements in many fields. Color texture segmentation has widespectrum of applications in diverse fields such as segmentation ofnatural images, medical image analysis, remote sensing, shapeextraction and inspection of products etc. This paper presentscolor texture segmentation algorithm which can satisfyrequirements for such applications. Proposed algorithm is basedon Markov Random Field (MRF) model eliminating the need ofmajor contributor viz. Gabor filter used in past four decades forfeature extraction and use only color as texture feature. Highlycrude segmentation results are produced using only color astexture features. Crude segmentation results are enhanced byusing Median filter with enlarged window size quantitativelydetermined by using parameters viz. structural similarity index(SSIM), mean square error (MSE) and peak signal to noise ratio(PSNR). Feature space dimensions are reduced by factor of 11 inproposed approach and this reduced computations by a factor of11. The experimentation is carried out on 80 multi-class colortexture benchmark images from Prague texture segmentationdataset and 4 benchmark images in Vistex dataset. Meansegmentation accuracy achieved for Prague texture dataset is87.55% and it is higher by 9.82% over the best performingalgorithm among 11 state-of-art algorithms suggested in mostrecent literature. Accuracy achieved for Vistex dataset is 98.21%.Average SSIM for Prague dataset is 0.91403 and Vistex dataset is0.9405.

School of Electronics & Communication Engineering

Engineering and Technology

World University root node

Fulltext

International Journal of Recent Technology and Engineering

Markov Random Field based Color Texture Segmentation

Texture segmentation is a well-known research domain and a large number of researchers are working on it across the globe due to its wide varieties of applications in various domains such as medical imaging, shape extraction, product inspection, remote sensing, and segmentation of natural images. Many times, images are corrupted by various types of noises such as Gaussian, speckle, and salt-pepper noise. Most often Gaussian noise is a source of corruption in many applications. Prague texture dataset is extensively used by researchers due to wide varieties of multi-class textures in it. The present study offers the restoration of Prague texture database benchmark images. These images are corrupted with Gaussian noise with different variance values. The state-of-the-art algorithm viz. Colour Block Matching 3D (C-BM3D) filter, which achieves the first rank among 15 algorithms reported in the most recent literature, is used for the restoration of noisy texture benchmark images. Image restoration performance metrics used are peak signal to noise ratio (PSNR) and structural similarity index (SSIM). © 2020 IEEE.

Proceedings of the International Conference on Intelligent Computing and Control Systems, ICICCS 2020

Quantitative Restoration of Noisy Colour Texture Segmentation Benchmark Images using State-of-the-Art Algorithm

Texture segmentation has a wide spectrum of applications in diverse fields. This paper presents an elaborated Fisher Linear Discriminant (FLD) based semi-supervised approach for improving the accuracy of segmentation of multi-class complex fine textures. Gabor filter and local statistics (local variance) are used for feature extraction of texture images. Texture segments in the image are separated using K-means clustering. The results obtained using K-means clustering are refined by multi-class Fisher Linear Discriminant (MFLD). The algorithm is tested on wide varieties of several hundred homogenous and complex textures from five texture databases viz. Outex texture database, vision texture database (Vistex), Brodatz textures, Prague textures and Pertex texture database. Fisher distance (FD) is a measure of texture separability. Segmentation of complex textures is relatively a difficult task. The improvement in the segmentation accuracy of complex textures is achieved simply by the termination of MFLD based algorithm when Fisher distance (FD) ceases to increase with the increasing iterations of MFLD. After a quantitative analysis of the experimentation, it is concluded that the segmentation accuracy of complex textures and the combination of complex and homogeneous fine textures (with small texture primitives) increases as high as 29.83% with the increasing iterations of MFLD resulting in a significant improvement at the boundaries. Detailed results are provided in the experimentation and results section of the paper. The results achieve the second rank for 21 benchmark images among the ten state-of-the-art algorithms.

Helix

A Multi-Class Fisher Linear Discriminant Approach for the Improvement in the Accuracy of Complex Texture Discrimination

The proposed paper presents a novel approach for classification of breast cancer. Digital mammograms have become the most effective techniques for the detection of breast cancer. The goal of this research is to increase the diagnostic accuracy of image processing and machine learning techniques for optimum classification between malignant and benign abnormalities in digital mammograms by reducing the number of misclassified cancers. The Local Binary Pattern method and Gabor Filter is used to extract the texture features of the digital mammographic images. The respective texture features are given to Support Vector Machine classifier which successfully classifies benign cases, malignant cases and normal breast cases with accuracy 96.72%, 84% and 81.90% respectively.

International Journal for Research in Applied Science & Engineering Technology

Classification of Breast cancer Using Local Binary Pattern and Gabor Filter

Image Segmentation is a very important aspect of perception and it still a very thought provoking subject for machine and computer perception. Studies in computer vision proves that the segmenting an image into various understandable regions for subsequent post processing (e.g., scene understanding, pattern recognition) is an ambitious task. This paper discusses the major colour image segmentation approaches. Then we review various algorithms with their advantages and disadvantages which try to increase class distinguishability. We are using the Berkeley Segmentation dataset BSD300 for segmentation which contains natural colour images. Lastly summarize the image segmentation methods, developed to integrate more feature information with high accuracy.

International Journal of Innovative Research in Science, Engineering and Technology

A Survey of Segmentation of Colored Images Based on Impressive Algorithms

Image Segmentation is an imperative part of computer perception and it stills a difficult and thriving subject for machine and computer perception. Research in computer vision proves that the segmenting of an image into various understandable regions for subsequent post processing (e.g., scene understanding, pattern recognition) is an ambitious task.In this paper, the proposed system uses various segmentation methods to increase the accuracy of segmentation and satisfactory visual entirety. The segmentation method is based on watershed, modified RAG-RKM and SLIC method. Using the Watershed segmentation we find how many pixels are allotted to which regions. Using modified RAGRKM method we find the mask and seed pixels using which the contour is plotted with the help of pixel value and angle of direction. The SLIC algorithm is used to find the segmentation of image using the angle of movement which takes into consideration the neighbourhood distance, edges of the regions and the area of the region. The proposed algorithm is applied on BSD300 database which contains 300 natural images. The accuracy of the proposed system is very high. Also the algorithm was applied on random color-texture images and the result found was good.

Segmentation of Colored Images based on Modified Region Merging on RAG-RKM together with SLIC

Breast cancer is second most dangerous disease in the world after the lung cancer among women. Because ofthis reason, breast cancer detection is most focused area by many researchers. Digital mammographic images andadvanced techniques are required by many researchers for developing Computer aided detection (CAD) for breastcancer detection. This study describes the recent advances in image processing and machine learning techniques forbreast cancer detection. The study shows that Local Binary Pattern method used for feature extraction and Supportvector machine for classification as foremost technique used for breast cancer detection. The comparative study ofliterature work summarizes the effectiveness of different approach used by researchers for breast cancer detection.

International Journal of Innovative Research in Electrical, Electronics, Instrumentation and Control Engineering

A Survey of Breast Cancer Detection and Classification Based on Texture Feature

This paper presents texture image segmentation based on the combination of phase correlation and region growing algorithm. Image segmentation is an integral part of image analysis. In image processing different processes are done on images to enhance it at different parameters. These processes are not directly applied on images but before that it needs to be segmented or divided in smaller parts called pixels or small blocks of pixels. So segmentation becomes an integral and basic part in image analysis and error at this stage can influence other processing techniques. As in this paper three algorithm are developed on two different dimensions for two class images. Phase correlation is used to find similarities in the windows of pixels which are of size 16x16. Similar and different pixel or pixel blocks are decided by the threshold value of the phase correlation result to find out which pixel falls into the same segments and which is not.

International Journal of Advanced Research in Electrical, Electronics and Instrumentation Engineering

Segmentation of Textured Images Using Region Growing and Phase Correlation Algorithm

Image segmentation for analysis is a major aspect of perception and till date it is challenging issue for machine perception. Many years of study in computer vision prove that segmenting an image into meaningful regions for subsequent processing (e.g., pattern recognition) is just as hard problem as invariant pattern recognition. In this paper, the proposed work uses the different type of procedures has been followed to carry out colour-texture image segmentation. Segmentation methods are designed to integrate more feature information, with high accuracy and satisfactory visual entirety. The segmentation process is based on MSST and student’s t-distribution method.