Graduate Seminar Series: Clinical Stream
Graduate Seminar Series for the Institute of Biomedical Engineering (BME). This day is for clinical stream presenters.
If you would like to invite your Principal Investigator, please add their email via the ‘Add Guest’ button and they will also be notified of your presentation.
Presentation Title: Enabling efficient information retrieval through a novel hybrid-architecture question-answer system and see-through, gaze-optimized keyboard
The most prevalent use of the internet is to connect individuals to the world through web search, which has become a principal means of human connection in contemporary times. While many people can control their muscles to use a computer or other internet-connected device, this is not always possible for individuals with significant physical impairments.
Human-computer interactions through natural language processing (NLP) may benefit individuals with significant physical impairments. Many of these individuals cannot control their muscles but can still communicate via eye movement and eye gaze. Eye-tracking was introduced more than a century ago and is now an established kinematic measurement. The recent integration of eye-tracking into virtual reality (VR) and augmented reality (AR) headsets makes eye-based control even more versatile.
This research aims to design and implement an alternative search engine to allow eye-gaze users to find answers to queries faster and more efficiently. A novel hybrid-architecture question and answer system will be programmed and deployed alongside an AR headset with embedded binocular eye-tracking. The overall system will comprise three components corresponding to the sequence of user-machine interaction. In the initial part, the user specifies the query through eye gaze on a novel persistent, see-through virtual keyboard. Next, by applying advanced NLP techniques, the system tries to understand the query and converses with the user to obtain the necessary information before generating an answer. Then, in the last part, the system retrieves data from the internet using search engine scraping techniques to generate the answer.
Twelve non-disabled adults will be recruited for two separate sessions. In first session, participants will gaze-type a specific passage in a timely manner while their eye movements will be tracked by the AR headset. The primary objective is to validate the proposed hardware setup and evaluate gaze typing performance with the proposed keyboard interface. In the second session, the feasibility of the hybrid-architecture question and answer system will be evaluated. The final study of this thesis will evaluate the system with 5 youth with severe physical impairments who currently use eye gaze as an access method.
Supervisor Name: Tom Chau
Year of Study: 3
Program of Study: PhD
Zoom link: https://us02web.zoom.us/j/89610372821?pwd=azd4SCtYVWtreVovaGNPV1c2NGY2Zz09
Meeting ID: 896 1037 2821
Powered by Calendly.com