A novel approach for medical video compression using kernel based meanshift ROI coding techniques
Trupti Barsakar, Vijay R. Mankar · 2016
Medical imaging is playing a vital role for development of medical facilities in rural areas. It has reduced the gap between patients and doctors, especially in case of distance hospitals for image diagnosis. But in case of emergency data transmission it is difficult to transfer a 10MB of images and videos. They can take half an hour to communicate even at higher speed. Many times minor part of the medical images frames might be diagnostically very useful, hence identification of Region of Interest technique is used for medical image and video compression. In this paper the proposed new technique is kernel based meanshift Region of Interest coding. Firstly this is applied for enabling superior image frames and examination of their weak localization. Secondly by applying lifting wavelet transform images can achieve precise image coded window based on Huffman Encoder. By implementing and analyzing this proposed work accurate region of interest added with non region of interest is accomplished and the final result gives us high resolution and lossless compression which is required for diagnosis of medical images as well as medical video compression. This result analysis mainly focuses on computing the Peak to Signal Noise Ratio, Mean Square Error and Compression Ratio of medical image frames.