UTAR Institutional Repository

Real-time face recognition social enhancement for visually impaired people

Tan, Jing Jie (2022) Real-time face recognition social enhancement for visually impaired people. Final Year Project, UTAR.

[img]
Preview
PDF
Download (11Mb) | Preview

    Abstract

    Vision is significant to every individual in their social life. Unfortunately, visually impaired people lose their sight and are only able to rely on the auditory sense to recognise the identity of a person. This makes visually impaired people (VIP) always be passive in social activity and affects their social confidence. However, there is a lack of affordable visual assistive devices to enhance their quality of social life. The novelty of this project is to develop a social enhancement assistant that assists VIP in both physical and virtual social activity. Hence, this project developed a face recognition mobile application that assists the VIP to identify a person from device cameras, external camera, and device’s screen. Meanwhile, this project also includes other recognition features such as emotion, gender, and age recognition. This application targets to improve VIP social life quality by being the eyes for them in meeting, presentation, chatting, and using social media. In fact, the fraud cases of misusing other’s identity towards VIP can also be reduced. Besides, the Machine Learning Kit Face Detection API and 3 light-weighted TensorFlow Lite face recognition models are implemented to achieve a high accuracy, and short recognition time. The model architectures are FaceNet – Inception-ResNet v1 (face identity), Mini-Xception (face emotion), and Inception module (face details). Besides, this application meets the requirement of TalkBack service in order to produce zero barrier for VIP when they are using this application. In addition, this application has visual accessibility user interface design and has implemented various interaction channels – vision, auditory, touch, vibrate and shake. Last but not least, this application integrated database backup functionality via Firebase Storage to ease data migration and ensure a minimum loss of face records.

    Item Type: Final Year Project / Dissertation / Thesis (Final Year Project)
    Subjects: Q Science > Q Science (General)
    T Technology > T Technology (General)
    Divisions: Faculty of Information and Communication Technology > Bachelor of Computer Science (Honours)
    Depositing User: ML Main Library
    Date Deposited: 15 Jan 2023 21:30
    Last Modified: 15 Jan 2023 21:30
    URI: http://eprints.utar.edu.my/id/eprint/4666

    Actions (login required)

    View Item