Deep Learning-Based Perceptual Stimulus Encoder for Bionic Vision

TL;DR

This paper introduces a deep learning-based perceptual stimulus encoder designed to improve the quality of artificial vision in retinal implants by predicting electrode activation patterns for desired visual percepts.

Contribution

It presents an end-to-end CNN model trained to generate electrode patterns for better visual perception in retinal prostheses, a novel approach in this field.

Findings

Effective on MNIST dataset

Validated with a psychophysical phosphene model

First step towards enhanced retinal implant vision

Abstract

Retinal implants have the potential to treat incurable blindness, yet the quality of the artificial vision they produce is still rudimentary. An outstanding challenge is identifying electrode activation patterns that lead to intelligible visual percepts (phosphenes). Here we propose a PSE based on CNN that is trained in an end-to-end fashion to predict the electrode activation patterns required to produce a desired visual percept. We demonstrate the effectiveness of the encoder on MNIST using a psychophysically validated phosphene model tailored to individual retinal implant users. The present work constitutes an essential first step towards improving the quality of the artificial vision provided by retinal implants.