Student research opportunities

Algorithms and scene representations to facilitate mobility with a Bionic Eye

Project Code: CECS_727

This project is available at the following levels:
CS single semester, Engn4200, Engn R&D, Honours, Summer Scholar, Masters, PhD

Keywords:

Computer Vision, Bionic Eyes, Navigation, Robotics, Scene Understanding, Segmentation

Supervisor:

Dr Chris McCarthy

Outline:

NICTA's "Vision processing for the Bionic Eye" (VIBE) project forms part of a national consortium, Bionic Vision Australia (BVA), seeking to develop a retinal implant for the blind. Vision is restored by transforming camera-captured pixel data into scene representations which are then rendered via electrical stimulation of retinal cells. A primary aim of VIBE (and BVA) is to deliver a system that will enable basic navigation for implantees. With a number of possible avenues to explore, you will be contributing to the development of vision algorithms and scene representations to support this.

Goals of this project

To gain familiarity with current 3D computer vision techniques, and the real-world challenges these techniques encounter.

To improve the effectiveness of current scene representations for navigation being examined (using our wearable prosthetic vision simulator).

AND/OR

The development of novel scene representations and/or underlying algorithms to support navigation with a low resolution vision prosthesis.

Requirements/Prerequisites

Basic image processing knowledge.
C/C++ programming

Student Gain

Excellent opportunity to be involved in a computer vision project dealing with real-world problems. Should the outcomes of your work offer improvements, there is the possibility of seeing your research forming part of the navigation system to be tested with implantees (within the next two years).

Links

NICTA VIBE project website
Bionic Vision Australia

Contact: