Integration of artificial vision with non-visual peripheral cues to guide mobility

Area of Science:

• Biomedical Engineering
• Computer Vision
• Neuroscience

Background:

• Visual prostheses offer limited vision restoration due to low resolution and restricted field-of-view.
• Severe vision loss presents challenges in navigation and environmental interaction.

Purpose of the Study:

• To develop and evaluate an advanced video processing system to augment existing visual prostheses.
• To improve navigation and task completion for individuals with severe vision loss.

Main Methods:

• A prototype system integrating a headworn depth camera and computer vision algorithms was developed.
• The system detects landmarks, plans paths, and provides vibrational cues for out-of-view targets.
• A case study was conducted with an Argus II retinal prosthesis user performing a navigation task.

Main Results:

• The participant's success rate in finding a door improved significantly with the augmented system.
• Task completion times were reduced with the integration of modified video and vibration feedback.
• The system demonstrated effectiveness in guiding users by leveraging both visual and non-visual sensory information.

Conclusions:

• Computer vision and non-visual cues can effectively enhance the functionality of visual prostheses.
• This approach offers a promising method for augmenting artificial vision and improving quality of life.
• Further development could lead to more sophisticated visual aids for the visually impaired.