Abstract
Purpose :
Luminance is a measure that describes the perceived brightness of a color, in this study we investigate the interaction between the color luminance of on-screen shapes and recognition accuracy as part of a home- based, rehabilitation program designed to help patients with bionic vision.
Methods :
Home-based computer rehabilitation modules were programmed using National Instruments Labview Development Suite (Austin, TX.). The patient is emailed a web link allowing the download and installation of the module. Module instructions are administered via computer synthesized voice, and subject is shown a random solid filled shape (circle, square triangle), colored (red, green, blue, white, gray, pink, purple, brown, orange, yellow) and centered on the screen. Each session has eight items, and there is no time limit during the viewing phase. The subject presses the spacebar to enter the recognition phase and selects the shape by keyboard number press (1, 2 or 3). At the completion of each module data measuring accuracy and timing are programmatically transmitted back for analysis via secure email. Two subjects, one implanted with the Argus II retinal prosthesis for 12 months, the other for 24 months, completed 58 sessions, recognizing 465 shapes.
Results :
Recognition accuracy was greater (82.6% ± 1.9) with low luminance colored shapes (brown, red, purple, and blue), compared to high luminance colors (gray, green, yellow, white) (76.0 ± 5.5), this difference trended toward statistical significance (p=0.09).
Conclusions :
Artificial vision devices provide ultra-low vision, allowing for crude perception of objects. High luminance colored objects may be more challenging to identify due to haloing of the contours, a consequence of limited brightness levels and low density of electrodes in the implanted array.
This is an abstract that was submitted for the 2017 ARVO Annual Meeting, held in Baltimore, MD, May 7-11, 2017.