Pre-processing technique for wireless capsule endoscopy image enhancement

Abstract

Wireless capsule endoscopy (WCE) is used to examine human digestive tract in order to detect abnormal areas. However, it has been a challenging task to detect an abnormal area such as bleeding due to poor quality and dark images of WCE. In this paper, pre-processing technique is introduced to ease classification of the bleeding area. Anisotropic contrast diffusion method is employed in our pre-processing technique as a contrast enhancement of the images. There is a drawback to the method proposed by B. Li earlier, in which the quality of WCE image is degraded when the number of iteration increases. To solve this problem, variance is employed in our proposed method. To further enhance WCE image, Discrete Cosine Transform is used with anisotropic contrast diffusion. Experimental results show that both proposed contrast enhancement algorithm and sharpening WCE image algorithm provide better performance compared with B. Li's algorithm since SDME and EBCM values are stable whenever the number of iterations increase. Moreover, the sharpness measurements using gradient and PSNR are both improved by 31.5% and 20.3%, respectively.