Real-time gastric intestinal metaplasia segmentation using a deep neural network designed for multiple imaging modes on high-resolution images

Passin Pornvoraphat, Kasenee Tiankanon, Rapat Pittayanon, Natawut Nupairoj, Peerapon Vateekul, Rungsun Rerknimitr

Abstract: Highlights•A real-time semantic segmentation network for gastric intestinal metaplasia (GIM) to prevent gastric cancer.•A modified deep neural network with a switch adapter module that supports multiple colour modes and different endoscope ranges in the GIM segmentation.•The model is enhanced by processing high-resolution images with a better view of the surface pattern.•Experiments reveal full comparisons of real-world scenarios and are validated by experienced endoscopists.