Image In painting and Compression Using DCT Algorithm

Abstract

This research work has major focus on images which have cracks and minor distortions in them. This prevents any compression technique to give the best output. Hence there is a need to first eliminate those distortions from the images before applying compression. A system is proposed, in which K-Nearest Neighbor Algorithm is implemented after the DCT Step, to remove the distortions before compressing it. The neighboring color of the distorted part can be used to merge with the damaged part. So basically, we have to store less color intensities to save the same image without the distorted piece which will result in more optimized compression ratio. It will be greatly valuable for observer perspective.