Influence of field of view in visual prostheses design: Analysis with a VR system

TL;DR

The paper addresses how field of view (FOV) and spatial resolution shape perceptual outcomes in retinal prostheses by introducing a panoramic VR-based simulated prosthetic vision system. It systematically varies circular FOVs ($20^{\circ}$, $40^{\circ}$, $60^{\circ}$) and phosphene counts ($200$, $500$) in a 360° panorama while subjects perform an object search/recognition task across 50 hotel-room scenes. The key finding is that higher angular resolution improves recognition accuracy and reduces response time, even as FOV narrows, with a diminishing return observed below $2.3$ phosphenes per degree, quantified through logarithmic relationships: $OR = -1.0345 + 0.4482 \cdot \log(AR)$ ($R^2 = 0.4031$) and $RT = 83.14 - 18.78 \cdot \log(AR)$ ($R^2 = 0.2344$). Overall, the study demonstrates that concentrating phosphene density to boost angular resolution is more beneficial than widening FOV, and it presents an open-source SPV platform for replication and extension in prosthesis design research.

Abstract

Visual prostheses are designed to restore partial functional vision in patients with total vision loss. Retinal visual prostheses provide limited capabilities as a result of low resolution, limited field of view and poor dynamic range. Understanding the influence of these parameters in the perception results can guide prostheses research and design. In this work, we evaluate the influence of field of view with respect to spatial resolution in visual prostheses, measuring the accuracy and response time in a search and recognition task. Twenty-four normally sighted participants were asked to find and recognize usual objects, such as furniture and home appliance in indoor room scenes. For the experiment, we use a new simulated prosthetic vision system that allows simple and effective experimentation. Our system uses a virtual-reality environment based on panoramic scenes. The simulator employs a head-mounted display which allows users to feel immersed in the scene by perceiving the entire scene all around. Our experiments use public image datasets and a commercial head-mounted display. We have also released the virtual-reality software for replicating and extending the experimentation. Results show that the accuracy and response time decrease when the field of view is increased. Furthermore, performance appears to be correlated with the angular resolution, but showing a diminishing return even with a resolution of less than 2.3 phosphenes per degree. Our results seem to indicate that, for the design of retinal prostheses, it is better to concentrate the phosphenes in a small area, to maximize the angular resolution, even if that implies sacrificing field of view.

Figures (9)

• Figure 1: Overview of a retinal prosthesis. The external and internal components include a glasses-mounted camera, an external processing unit and an implanted electrode array. First, the external camera acquires an image. Then, the external processor converts the image to a suitable pattern of electrical stimulation of the retina through an electrode array. The result is a phosphene image with limited field of view (FOV).
• Figure 2: SPV system. The components consist of an Oculus Rift powered by a consumer level laptop. The VR system is composed by two lenses, two screens and a suite of internal sensors (gyroscope and accelerometer). The representation with simulated phosphenes is displayed on the laptop screen as well as on the Oculus system worn by the subjects. During the experiment, subjects were seated in a swivel chair allowing them to scan the entire scene all around them (360 degrees).
• Figure 3: Data process. The input scene in our VR system is a panoramic image in equirectangular representation. The system estimate the head orientation using the IMU sensors (gyroscope and accelerometer) allowing to choose the area of the panoramic scene that is being observed at the moment. The area selected is then projected on the two Oculus screens and represented using simulated phosphenes.
• Figure 4: Stimuli conditions in the experiment. Rows: different circular FOVs used in the experiment (60, 40 and 20 degrees). Columns: different resolutions used in the experiment (200 and 500 phosphenes). Note that in the last row only part of the lamp is visible, the beds cannot be seen without moving the point of view.
• Figure 5: Object classes considered during the experiment. Subjects have to recognize the main objects of the scene, those objects that are usually present in hotel rooms such as bed, window, chair, tv, sofa, table/nightstand, door, wardrobe/shelving and lamp.