Recognition of Humans using Simulated Prosthetic Vision
Keywords
Loading...
Authors
Issue Date
2021-07-01
Language
en
Document type
Journal Title
Journal ISSN
Volume Title
Publisher
Title
ISSN
Volume
Issue
Startpage
Endpage
DOI
Abstract
Focusing on one of the most signi cant stimulus types that we encounter in our daily lives, we
present and compare several methods to represent freely moving humans in prosthetic phosphene
vision via realistic simulations and behavioural experiments. The results show that AI-powered
methods only out-perform the baseline (edge detection) when classifying multiple people. Be-
sides that, two of the three proposed representations used signi cantly less phosphenes than the
baseline which is bene cial in real life scenarios. Although there is much room for further im-
provement, this experiment is a proof of concept that AI-powered representations can enhance
the vision generated by VCPs.
Description
Citation
Supervisor
Faculty
Faculteit der Sociale Wetenschappen