Detecting gastric cancer from video images using convolutional neural networks

Abstract

Early detection of gastric cancer is one of the most important factors for improving the prognosis of patients. However, the detection rate for gastric cancer differs depending on the endoscopist's experience. In the last decade, new deep‐learning‐based machine learning methods have shown significant improvements in image recognition and have therefore been applied to various medical fields. Previously, we have reported the efficacy of our deep learning‐based convolutional neural network (CNN) system for detecting gastric cancer in still images. This CNN system achieved a high detection rate (overall: 92.2%, diameter greater than 6 mm: 98.6%) with a high processing speed (2296 images/47 sec.) [5], and it is expected to help endoscopists increase diagnostic yields.

This article is protected by copyright. All rights reserved.

https://ift.tt/2FusMXr