An automatic blood detection algorithm for wireless capsule endoscopy images

Wireless Capsule Endoscopy (WCE) is a noninvasive color imaging technique that has been introduced for the screening of the small intestine, which is difficult to reach through conventional endoscopies. A major problem with this technique is that it generates a huge number of images, whose analysis is very time consuming and tedious for physicians. To help the physicians, we propose an automatic blood detection algorithm for WCE images. Utilizing a particular Ohta color channel (the average of the red, green, and blue channels of the input image), we employ analysis of eigenvalues of the image Hessian matrix and multiscale image analysis approach, for designing a function to discriminate between blood and normal frames. Experiments are performed on our current data set to verify the proposed algorithm.