Correction of Perceived Visual Distortions and Correlation to Retinal Changes in Eyes with Age-Related Macular Degeneration
Advisor Information
Mahadevan Subramaniam
Location
Dr. C.C. and Mabel L. Criss Library
Presentation Type
Poster
Start Date
3-3-2017 2:15 PM
End Date
3-3-2017 3:30 PM
Abstract
AMD causes vision loss in millions due to permanently damaged photoreceptors in the macula of the eye, where the sharpest central vision occurs. Visual distortions due to damaged photoreceptors and imprecise positioning of retinal implant treatments make it difficult to perform daily activities. Our objective is to develop a wearable device that can automatically measure distortions perceived by AMD subjects and automatically correct these distortions to restore functional vision in AMD patients. Estimation and the automated correction of visual distortions by our learning algorithms are highly promising in restoring functional vision to AMD patients. Our preliminary Phase-1 trials using our software app achieved around 75% improvement in distorted vision in all patients. Distortions measured in the trials were independently verified by multiple graders and show high correlation to the affected macular regions witnessed in the OCT images. Our image processing software that automatically superimposed the distortions from the software app to the OCT images achieved significant correlation (p < 0.05) between visual function and diseased macular regions. We are developing an improved version of the app to correct higher order visual activities such as reading and ambulation. We are enhancing our machine learning algorithms to dynamically adapt for eye movements and for the error in distortion measurement due to individual variances.
Correction of Perceived Visual Distortions and Correlation to Retinal Changes in Eyes with Age-Related Macular Degeneration
Dr. C.C. and Mabel L. Criss Library
AMD causes vision loss in millions due to permanently damaged photoreceptors in the macula of the eye, where the sharpest central vision occurs. Visual distortions due to damaged photoreceptors and imprecise positioning of retinal implant treatments make it difficult to perform daily activities. Our objective is to develop a wearable device that can automatically measure distortions perceived by AMD subjects and automatically correct these distortions to restore functional vision in AMD patients. Estimation and the automated correction of visual distortions by our learning algorithms are highly promising in restoring functional vision to AMD patients. Our preliminary Phase-1 trials using our software app achieved around 75% improvement in distorted vision in all patients. Distortions measured in the trials were independently verified by multiple graders and show high correlation to the affected macular regions witnessed in the OCT images. Our image processing software that automatically superimposed the distortions from the software app to the OCT images achieved significant correlation (p < 0.05) between visual function and diseased macular regions. We are developing an improved version of the app to correct higher order visual activities such as reading and ambulation. We are enhancing our machine learning algorithms to dynamically adapt for eye movements and for the error in distortion measurement due to individual variances.