The following is a summary of “Real-world indoor mobility with simulated prosthetic vision: The benefits and feasibility of contour-based scene simplification at different phosphene resolutions,” published in the February 2022 issue of Ophthalmology by Steveninck, et al.
The use of neuroprosthetic implants as a potential solution for restoring some level of vision in people with visual impairments was a promising field of research. While the perception generated by these devices was still limited compared to normal vision, it could provide essential information for basic daily tasks such as mobility. Several studies have investigated the benefits of visual neuroprosthetics in simulated prosthetic vision paradigms, particularly concerning scene simplification through image processing, such as contour extraction.
For a simulation study with sighted participants, researchers sought to explore the attainable benefits of scene simplification in an indoor environment with varying environmental complexity and the practical improvement achieved with a deep learning-based surface boundary detection implementation compared to traditional edge detection.
The study found that a simulated electrode resolution of 26 × 26 was sufficient for mobility in a simple environment. Additionally, removing background textures and within-surface gradients may improve mobility for a lower number of implanted electrodes. However, the current study’s deep learning-based implementation for surface boundary detection did not improve mobility performance. Moreover, removing within-surface gradients and background textures may worsen mobility performance for a greater number of electrodes. Thus, determining the optimal level of scene simplification depends on the number of implanted electrodes and required a careful balance between informativity and interpretability.