Active Confocal Imaging System for Visual Prostheses
Abstract
Objectives and Rationale: Blindness is devastating, significantly affecting one s quality of life and adding a support burden for caregivers and family members. In recent years, there have been great advances in prosthetic vision. Among the most exciting is the recently Food and Drug Administration-approved Argus II retinal implant. In addition to retinal implants, there are numerous efforts to create devices that will substitute other senses for vision. These devices, called "sensory substitution devices" (SSDs), use various combinations of audio, tactile, or electrical (tongue) stimulation. However, these all suffer from low resolution, low dynamic range, and limited visual field. Even with the most optimistic projections, these limitations are likely to remain. Low resolution results in an "image" that is in the form of dots, as if looking at a picture in a "connect-the-dots" children s book. Low dynamic range results in a picture that is in only a few shades of gray -- sometimes even just black and white. Limited visual field means that only a small fraction of a scene is visible at one time. These limitations currently make these devices only slightly better than useless. For example, reading using the Argus II requires tens of seconds for single letters and minutes for short words. At these rates, interpreting an image or a scene while walking is almost impossible. In addition, these rates were measured only in "clean" lab settings without the clutter that plagues real-world images. (Think lots of dots in addition to those needing connection.) Our proposed system will make it easier to use these devices. In processing a scene, it will automatically identify a few relevant distinct objects at different distances. The user can then manually select the distance/object of interest from the few automatically selected ones. Our device will remove background clutter (objects at other distances), which will make it easier to recognize objects of interest with prosthetic vision. Our system will employ new technology called "light-field" or "integral imaging" that can focus on objects at one selected depth plane (distance of object of interest) at a time (after the photo was taken), and importantly, blur and thus suppress the clutter from objects at other depth planes. We propose to build a prototype and validate the system using a low-resolution head-mounted display (HMD) and the tongue stimulation "BrainPort" device. We will test first with readily available normally sighted subjects subjects, simulating phosphene prosthetic vision in the HMD. We will also train 20 blind subjects in the use of the BrainPort tongue stimulator. Clinical testing will be composed of three studies. All studies begin with simulated phosphenes in the HMD, then the BrainPort, and then available patients implanted with the Argus II. Study 1 will measure object recognition on static images, Study 2 will test object search on static images, and Study 3 will test mobility while traversing an obstacle course. Ultimate Applicability and Potential Impact: Ultimately, these techniques of preparing the image will make retinal implants and other prosthetic vision devices much more useful. Doing so will have a major impact on the restoration of vision. By mitigating the limitations of current and projected prosthetics, our techniques extend the applicability and usefulness of prosthetic vision for the rehabilitation of debilitating vision loss. Potential Clinical Applications, Benefits, and Risks: As the front end to any visual prostheses envisioned today, it will substantially improve the user s performance by translating basic novel optical technology into a vision rehabilitation technology. The risk is minimal, since this will be an "add-on" to existing and future devices. The benefits will be to make these devices much more useful. Types of Patients Will It Help, and How Will It Help Them: This device will help any
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Apr 04, 2016
- Source ID
- W81XWH1610033
Entities
People
- Eli Peli
Organizations
- Schepens Eye Research Institute
- United States Army