Detection of Gastrointestinal Disease in Endoscope Imaging System

摘要

Conventionally, endoscopy produced a scanning result video which provides information in medical diagnosis. The advanced in computer and image processing technology have allowed the video captured being converted into series of images for further processing. Therefore, the focus of this research is to help physician in gastrointestinal disease detection by providing assistant tool in diagnosis process. By using structural similarity technique as pre-processing, image normalization is used to enhance the image quality, and use the colour feature analysis as the detecting tool. As the result, the lengthy endoscopic video can be reduced up to 70% and the system developed has the capability to highlight the suspected area as the reminder for physician in diagnosis.