Skip to content

Latest commit

 

History

History
53 lines (29 loc) · 2.03 KB

README.md

File metadata and controls

53 lines (29 loc) · 2.03 KB

G-Meet

On-line non-face-to-face Exam platform using WebRTC & Custom YOLO

Video link : youtube


Abstract

As face-to-face classes have recently become difficult due to COVID-19, most tests are conducted non-face-to-face. all students taking exams on video conferencing platforms such as Zoom and Webex, which are popular today, adjust their cameras so that their face and monitor hands are visible on camera and take the exam.

However, there are many difficulties in actually conducting non-face-to-face test supervision using the above platform.

To address many issues, we devised an easy way to supervise cheating while going off-line. we have implemented a platform that helps you easily manage a large number of students when taking non-face-to-face exams by combining AI differently from existing general video meeting platforms. Online meeting platform is built using WebRTC, a real-time communication method that provides a video stream so that students can check their webcams or monitor screens.

Added the ability to monitor for cheating.

Features

  • On-line testing platform using his own web browser without any plugins
  • Share his own media stream to other peers without intermediate server
  • Object detection model that determines whether user is cheating and plots detection result graph

Demo

Main browser

  • Login browser loginScreen

Test taker's browser

testtaker

Supervisor's browser

supervisor

YOLO custom model

model1

model2

Author