Improving and optimizing video contrast for endoscopy imaging

In this paper, we present an endoscopy video contrast improvement optimization method that provides contrast enhancement without noise perceptible increase. The method is based on video data clustering of fine picture details (edges and textures) and of large details. The texture class contains only details that can be detected by visual perception. This class includes the detected texture and picture noise, and should not be subjected to contrast enhancement since it does not reduce video quality. Thus it is possible to prevent noticeable increase in noise through video contrast improvement. The experimental results of endoscopy video contrast enhancement are discussed in this paper.