New color image enhancement method for endoscopic images

Abstract

This paper presents a simple and efficient color image enhancement method for endoscopic images. The proposed image enhancement method works by two interrelated steps: image enhancement at gray level and color reproduction. At, first, the captured RGB endoscopic image is converted into 2 dimensional gray level spectral images using a well-known method called FICE (Fuji Intelligent Color Enhancement). In the next stage the image with maximum entropy is selected as the base image to be used for color reproduction. Maximum entropy value indicates the maximum enhanced image. In color reproduction, the entire chrominance map of a source image is transferred into the base image by matching luminance and texture information between two images. The distance between luminance components of target (base image) and source images are calculated using 2-norm Euclidean distance. The proposed color image enhancement method is compared with popular narrow-band imaging on the scale of image quality, image enhancement, simulation speed and efficiency of color reproduction and distortion. The proposed method can be applied on any RGB images collected from any white light endoscopic devices. Is highlights the tissue characterization on the surface part of base endoscopic image that enables better diagnosis.