We propose to develop and evaluate the utility of a novel input camera system for retinal prosthetics that allows users to filter optical imagery based on distance from the observer. We call this depth filtering, which can be accomplished with a depth sensing stereoscopic camera system. Retinal prostheses can restore partial vision to people blinded by outer retinal degenerative diseases such as Retinitis Pigmentosa (RP) or Macular Degeneration. The Argus II Retinal Prosthesis System is intended to provide electrical stimulation of the retina to elicit visual perception in blind individuals with (near) total visionloss due to retinitis pigmentosa. The implanted epiretinal array provides a 10 x 6 grid of electrodes. Electrical pulses at these 60 pixel sites stimulate the retina's remaining cells and result in the perception of patterns of light. The current spatial resolution is only 60 total pixels, but the camera systems that provide image information to the prosthesis user transmit full-resolution video information. Information in the imagery consists of all visible items within the field of vie ranging for example from nearby obstacles to far away mountain scenery. The lack of perceived spatial resolution makes it extremely difficult for prosthesis users to differentiate between important nearby elements within the scene and relatively unimportant distant elements of the scenery, or vice versa. An image capture and processing system is proposed that selectively removes elements in a viewed scene that are outside a user-set depth range. We hypothesize that system utility in important tasks such as people or object finding and mobility will be less burden some, and performance improved, by providing only relevant visual information to the wearer. The proposed project seeks to develop and experiment with distance-filtered optical imagery as an input modality to retinal implants. Two prototype systems will be specified, designed, built and evaluated in Argus II patients: In phase 1 of the proposed project the image processing will be performed by a computer external to the Argus II system. In phase 2, a more powerful version of the next-generation Argus II video processing unit, capable of performing the image processing, will be utilized. This will be coupled with a new dual camera eyeglass assembly, making the phase 2 prototype fully integrated and ready for commercialization in the next generation Argus II system, and possibly in other visual prosthesis systems and head-mounted electronic low vision aids.
Retinal prostheses aim to help people who have lost their vision due to degenerative eye conditions such as retinitis pigmentosa to perceive shapes and movement, to become more mobile, or to perform other day-to-day activities. The proposed project relates to the utilization of distance-specific optical imaging as an input modality to a retinal prosthesis, thus improving the prosthesis wearer's understanding of the scene by removing unwanted/irrelevant information.
