Semantic Gait Parameter Extraction from Video Data Using Generative Models

This study addresses the challenge faced by children with severe motor impairments in acquiring proficient powered wheelchair (PW) skills for independent locomotion. Traditional means, such as joysticks, may not be suitable, prompting exploration of brain-computer interface (BCI)-based PWs. However, limited research exists in the pediatric realm due to developing brains and a lack of tailored BCI systems. Among BCI control schemes, the motor imagery (MI) paradigm shows promise for real-world use as it doesn’t rely on visual stimuli. Despite requiring focused attention and causing fatigue, MI competes with the more user-friendly P300, which boasts high accuracy and ease of use. The study aims to compare the usability of MI and visual P300 paradigms for controlling PWs in real contexts, considering the perspectives of parents, service providers, and observations. By delving into this unexplored area, the research seeks to enhance assistive technology for children with severe motor disabilities, contributing valuable insights to both home and host institutions. This project aligns with social responsibility goals, promising advancements in knowledge and technology to improve the lives and mobility of vulnerable populations. Overall, participating institutions stand to benefit by contributing to knowledge, refining assistive technology, and addressing the genuine needs of affected children.

Faculty Supervisor:

Ervin Sejdic

Student:

Partner:

Flinders University

Discipline:

Engineering

Sector:

Education

University:

University of Toronto

Program: