DEVELOPMENT OF A DEEP LEARNING TECHNOLOGY TO IMPROVE HUMAN-COMPUTER INTERACTION FOR VISUALLY IMPAIRED

Subject Area: Computer Science

Accessing equitable learning settings can be extremely difficult for people with disabilities, such as vision loss, despite the fact that receiving a high-quality education is a fundamental right. The main goal of this study is to improve accessibility in educational settings by developing an intuitive system that combines deep learning technology to improve human-computer interaction for visually impaired users. The study's primary goal was to create a biometric user identification system that uses facial recognition to give students safe, customised access. A robust object identification system that can reliably identify people in real-time was demonstrated using You Only Look Once (YOLOv5), enabling quick login procedures based on face biometrics. A better human-computer interaction platform that caters to the unique requirements of visually impaired users was also modelled in addition to the authentication system. By combining speech-to-text and voice commands, this technology enables students to freely explore and interact with a computer interface while taking tests. An adaptive ambient noise cancellation technique that uses the Least Mean Squares (LMS) filter to cut down on background noise was developed as a way to improve this system even further. For real-time human-computer interaction, the LMS filter continually adjusts to shifting noise levels, enhancing voice recognition performance and intelligibility in loud settings. Lastly, the study combines these models into a single human-computer interactive system that combines adaptive noise cancellation, voice recognition, and biometric identification. It then deploys the system for analysis. The system's efficacy in offering a user-friendly, accessible solution for visually impaired people in educational settings was confirmed by evaluating its performance using real-world test scenarios. The study was recommended for deployment at school and colleges where students with impaired vision are admitted to help improve their quality of education.