Thesis: Bionic eye: Development of a neuromorphic system for processing visual input

Gáspár, N. (2015) Thesis: Bionic eye: Development of a neuromorphic system for processing visual input. Master's Thesis / Essay, Biology.

Preview	Text MasterLS_BME_2015_NGaspar.pdf - Published Version Download (1MB) | Preview
	Text Toestemming.pdf - Other Restricted to Backend only Download (66kB)

Abstract

There are approximately 285 million blind or visually impaired people worldwide. Even though in the last couple of years there were remarkable improvements in the field of Retinal Prosthetics, there are still plenty of opportunities for further development. During this project, a novel, real-time visual to auditory sensory substitution system (SSD) was created, mimicking the information processing strategy of the retina. With few small alterations, the same system could be also used as a front end of a retinal implant. The typical architecture of a Retinal Prosthetic system uses the following overall strategy: The image is captured by a visual sensor, and forwarded to be processed. The output data is then forwarded to a stimulator used to induce signals that are interpreted by the brain. In the case of implants, the stimulation signal is a small electrical impulse, while in the case of SSD devices, sound or vibro-tactile feedback are used for stimulation. In this project, instead of a conventional camera, an embedded Dynamic Vision Sensor (eDVS) was used. While conventional cameras capture inputs as a series of frames, a DVS sensor only transmits pixel-level changes, this way simulating the behaviour of the retinal cells. The input from the camera is forwarded to a microcontroller. A simple algorithm based on an Integrate & Fire neuron model was used to emulate Retinal Ganglion Cells (RGCs), and a mapping algorithm was applied to convert visual information to sound, in a 4x4 coordinate system. The stimulator array of the possible retinal implant was represented with an 8x8 RGB LED array. The stimulator of the SSD device consisted of an Arduino board and a custom built electronic circuit. As a final step, the same visual to audio conversion algorithm was implemented as an Android application. The main scope of this project was to develop a proof-of-concept system. Several possibilities for future improvements in terms of image resolution and user experience are discussed.

Item Type:	Thesis (Master's Thesis / Essay)
Degree programme:	Biology
Thesis type:	Master's Thesis / Essay
Language:	English
Date Deposited:	15 Feb 2018 08:06
Last Modified:	15 Feb 2018 08:06
URI:	https://fse.studenttheses.ub.rug.nl/id/eprint/13128

Actions (login required)

View Item