Project Overview

Title: Video Pipeline for Machine Computer Vision

Client: JR Spidell

Advisor: Dr. Zambreno and Dr. Jones

Description: Pupil detection subsystem using machine learning algorithms on an FPGA to support our client's vision of advanced assistive technologies.

Highlighted Technologies: OpenCV, PyTorch, Vitis-AI, Tensil-AI, and the Ultra96v2

Problem Statement

People with mobility and cognitive impairments, such as Cerebral Palsy, face significant challenges in maintaining independence and safety. Traditional wheelchairs often lack the advanced technologies needed to support these users, leaving gaps in autonomy, communication, and safety. Healthcare professionals and caregivers also struggle with the absence of real-time alerts for medical emergencies like seizures, increasing the risk of delayed responses. These challenges not only affect the quality of life for wheelchair-bound individuals but also limit opportunities for proactive care.

Our client wants to address these issues by developing assistive wheelchair technologies with features such as advanced mobility assistance and real-time seizure detection. This system aims to increase wheelchair user autonomy, improve safety, and reduce caregiver stress. Our team is collaborating with the client to develop a subsystem that detects, locates, and presents information on the user’s eyes in real time that will be used in future iterations of the client’s vision.

Overall System Flow: