National Taiwan University

Li-Yang Chang

Chung-Ping Chen

Yi-Hsuan Chou

Han-Mo Chiu

Shuo-Hong Hung

A system for determining a colonoscopy image is disclosed. The system includes a bowel preparation evaluation unit, an image processing unit electrically or communicatively connected to the bowel evaluation unit, and a convolutional neural network unit electrically or communicatively connected to the image processing unit. The bowel preparation evaluation unit determines whether a bowel in the colonoscopy image is clean. The image processing unit receives the colonoscopy image determined to have a clean bowel and outputted by the bowel preparation evaluation unit, and performs an image process on the colonoscopy image with clean bowel to adjust a contrast thereof. The convolutional neural network unit comprises a convolutional neural network and in a working mode, receives and determines whether the processed colonoscopy image is a cecal image, a non-cecal image, or an indeterminable colonoscopy image.

Growing community of inventors

Taipei, Taiwan

Li-Yang Chang