AutoNotes: A Touch-Free Blink-Based Interactive Model for Generation of Notes from Lecture Videos

ABSTRACT

The recent explosion of interest in online courses can directly be attributed to the current pandemic which has collated to the adoption of Massive Open Online Courses. While the enforced social distancing protocol inevitably demands the involvement of technology in raising awareness, it further requires the imposition of preventive restrictions on touch-based systems. Visual gestures like simple blinks can account for a significant subset of possible actions in mobile application domains. The proposed model introduces a novel approach that facilitates the automated generation of relevant lecture notes and web-linked keywords from videos through a blink-controlled interface. Functions like auto-video pausing and auto-closing of the application based on eye attribute tracking have been integrated into the system for handling context switching. The experimental results reveal 84.17 usability score, an average of 89.55% of different types of blink detection accuracy, 91.56% of text generation accuracy and 70.41% of keyword detection accuracy along with low false positives and false negatives under different lighting conditions. © 2021, The Institution of Engineers (India).