K. K. Bhoyar

Work place: Yeshwantrao Chavan College of Engineering Nagpur, 441110, Maharashtra, India



Research Interests: Mathematics of Computing, Image Processing, Computer systems and computational processes, Computational Science and Engineering


Dr. K.K. Bhoyar completed B.E. in Computer Science and Engineering in 1990 and M.Tech. in Computer Technology in 2001, both from SGGSIE&T Nanded , an autonomous Institution funded by Govt. of India. He has been awarded Ph.D. degree in Computer Science & Engg. from VNIT , Nagpur in July 2010, for his research work titled ‗Performance Enhancement of Color Based Classification, and Segmentation for Image Retrieval using JND Approach.‘ He is presently working as Professor, in the Dept. of Information Technology at Yeshwantrao Chavan College of Engineering Nagpur (India). His areas of interests include Image Processing and Soft Computing.

Author Articles
Image Classification Using Fusion of Holistic Visual Descriptions

By Padmavati Shrivastava K. K. Bhoyar A.S. Zadgaonkar

DOI: https://doi.org/10.5815/ijigsp.2016.08.07, Pub. Date: 8 Aug. 2016

An efficient approach for scene classification is necessary for automatically labeling an image as well as for retrieval of desired images from large scale repositories. In this paper machine learning and computer vision techniques have been applied for scene classification. The system is based on feature fusion method with holistic visual color, texture and edge descriptors. Color moments, Color Coherence Vector, Color Auto Correlogram, GLCM, Daubechies Wavelets, Gabor filters and MPEG-7 Edge Direction Histogram have been used in the proposed system to find the best combination of features for this problem. Two state-of-the-art soft computing machine learning techniques: Support vector machine (SVM) and Artificial Neural Networks have been used to classify scene images into meaningful categories. The benchmarked Oliva-Torralba dataset has been used in this research. We report satisfactory categorization performances on a large data set of eight categories of 2688 complex, natural and urban scenes. Using a set of exhaustive experiments our proposed system has achieved classification accuracy as high as 92.5% for natural scenes (OT4) and as high as 86.4% for mixed scene categories (OT8). We also evaluate the system performance by predictive accuracy measures namely sensitivity, specificity, F-score and kappa statistic.

[...] Read more.
Other Articles