PROJECT TITLE :

sequence-to-sequence similarity-based filter for image denoising - 2016

ABSTRACT:

Image denoising has been a well-studied drawback for imaging systems, especially imaging sensors. Despite outstanding progress in the standard of denoising algorithms, persistent challenges remain for a large class of general pictures. In this paper, we have a tendency to present a replacement concept of sequence-to-sequence similarity (SSS). This similarity live is an economical methodology to judge the content similarity for images, particularly for edge information. The approach differs from the ancient Image Processing techniques, which depend upon pixel and block similarity. Based mostly on this new concept, we introduce a new SSS-based filter for image denoising. The new SSS-based mostly filter utilizes the edge data in the corrupted image to address image denoising problems. We have a tendency to demonstrate the filter by incorporating it into a replacement SSS-based mostly image denoising algorithm to get rid of Gaussian noise. Experiments are performed over artificial and experimental knowledge. The performance of our methodology is experimentally verified on a selection of pictures and Gaussian noise levels. The results demonstrate that the proposed method's performance exceeds several current state-of-the-art works, which are evaluated both visually and quantitatively. The presented framework unveil new views in the use of SSS methodologies for Image Processing applications to switch the ancient pixel-to-pixel similarity or block-to-block similarity.