Skip to main content
. 2025 Mar 24;25:144. doi: 10.1186/s12911-025-02973-1

Table 4.

Real-time prediction of videos by the efficientnet-hill model

Video Links
Online Streaming Link

Video 1: https://share.weiyun.com/mGY6cCHW

Video 2: https://share.weiyun.com/IDswqAk4

Scan QR Code to Watch Video Online graphic file with name 12911_2025_2973_Tab4_HTML.gif

This video series (including Video 1 and Video 2) demonstrates the model's effectiveness in real-time Hill classification of gastroesophageal flap valves. In Video 1, the top left corner displays the model's predictions for the top two categories and their corresponding confidence levels in red font, while a real-time dynamic heatmap reveals the key areas of focus for the model. Video 2 showcases the process of real-time inferential prediction on multiple GEFV images displayed on an iPad, using a camera after the model is deployed on a local computer