Application Progress of Artificial Intelligence Technology in Image Recognition of Capsule Endoscopy / 胃肠病学

ABSTRACT

Capsule endoscopy (CE) is the main method to detect small intestinal lesions. However, a single CE examination produces about 60 000 images, to screen lesion from the huge amount of images is a time-consuming, boring work, and is easy to cause missed diagnosis because of the limited experience and professional skill of physician. Therefore, it is urgent to develop a system that can automatically detect intestinal lesions. In recent years, the technique of artificial intelligence (AI) has gradually penetrated into the medical field, and the computer-aided diagnostic technology based on big data and cloud computing has become a hot spot of clinical research. The deep learning (DL) model represented by convolutional neural network (CNN) has the ability of rapid recognition of lesions and can effectively reduce the missed diagnosis rate. This article reviewed the application progress of AI technology in image recognition of CE.