Real-time Computer Vision
Supervisor
Suitable for
Abstract
Co-supervised by Julian Godding (j.godding@gardin.co.uk)
This project will discover novel methods to perform high-accuracy computer vision tasks on streamed camera images (>10 fps) using a raspberryPi CPU. This presents challenges due to the accuracy/compute trade-off. The student would need to test several different architectures, design novel solutions for online training and assess performance on a bespoke data set.