Add-on optical mask to declutter visual information based on depth for visual prostheses

Area of Science:

• Biomedical Engineering
• Computer Vision
• Optical Physics

Background:

• Visual prostheses often struggle with scene information decluttering, leading to perception issues where bright, distant objects obscure nearby obstacles.
• Current depth sensing methods like Lidar and stereo imaging are unsuitable for wearable visual prostheses due to power or computational demands.

Purpose of the Study:

• To develop a simple, passive, and robust depth sensor for visual prostheses.
• To enhance the delivery of relevant visual information to implants by incorporating distance data.

Main Methods:

• An optical mask was designed and fabricated to induce chromatic aberrations and wavelength-dependent focal plane shifts.
• When placed in front of an RGB camera, the mask causes different colors (red, green, blue) to focus at distinct distances (near, mid, far).
• A standard deviation image filter was employed to extract in-focus information and convert color data into depth information.

Main Results:

• The optical mask successfully enabled depth perception by separating focal planes based on color wavelengths.
• Red, green, and blue channels focused on near, mid, and distant objects, respectively.
• The system effectively converted color-coded depth information into a usable format.

Conclusions:

• The proposed optical mask provides a viable solution for passive depth sensing in visual prostheses.
• This technology can significantly improve the functionality and usability of visual prostheses, particularly in outdoor environments.
• The miniature, passive nature of the device makes it suitable for integration into wearable visual aid systems.