Face Recognition Based Video Conferencing Add-On for Online Session Log Using Convolutional Neural Networks

Video conferencing is used to conduct sessions during pandemics and there is a need to maintain logs for assessment and evaluation purposes. In this work, an add-on tool for attendees data collection is developed. The system helps to mark session attendance of participants. The HOG and Face landmark estimation algorithms are used over SSVM, ANN, and CNN classifiers to assess the face recognition in real-time. The system captures faces from the video frames and compares them with the dataset images. System matches the face registered and marks the attendance in the datasheet. The system is initially trained and tested with the unique faces of the attendee, which in turn creates the database. A session log report is generated. This extension add-on tool is simple and useful for the education system to create logs without much effort and session time can be utilized for quality content delivery.