Abstract
Purpose: :
To provide an objective, subject-independent means for testing and validating the efficiency and utility of real-time image processing algorithms for enhanced visual perception and independent mobility for the blind using visual prostheses.
Methods: :
Testing time with visual prosthesis carriers is at a premium. Thus, it is helpful to have a more realistic functional approximation of a blind subject. Instead of a normal subject with a healthy retina looking at a low-resolution (pixelated) image on a computer monitor or head-mounted display, a better approximation is achieved by employing an objective, subject-independent mobile robotic platform that uses a pixelated view as its sole visual input for navigation purposes. Such a mobile robotic platform represents a constantly available testbed for real-time (e.g., 30 fps) image processing systems such as the Artificial Vision Simulator (AVS; Fink and Tarbell, ARVO 2005). AVS provides a wide variety of image processing modules and the capability to engage these repeatedly in any user-defined order to enhance the visual perception afforded by visual prostheses.
Results: :
We have created CYCLOPS, an AWD remote controllable robotic platform with wireless Internet connectivity and a fully articulated digital camera with wireless video link. CYCLOPS supports both interactive telecommanding via joystick, and autonomous self commanding. Due to its powerful onboard computing capabilities and long battery life, CYCLOPS can perform complex and numerically intensive calculations, such as: (1) testing and validation of image processing algorithms (e.g., AVS) that may enhance the visual experience afforded by visual prostheses; (2) testing of navigation algorithms and strategies to improve the degree of unaided mobility; and (3) testing of additional sensors (e.g., infrared) to further enhance the quality of life for blind subjects using visual prostheses. CYCLOPS uses only the pixelated camera images to move about an environment (e.g., room/corridor with obstacles) by processing them via AVS to result in new motion and navigation commands (e.g., "find the open door in a room").
Conclusions: :
Its Internet connectivity renders CYCLOPS a worldwide accessible testbed for researchers in the field of artificial vision systems. CYCLOPS enables objective, subject-independent evaluation and validation of AVS-type image processing systems with respect to their utility and efficiency to support and enhance visual prostheses, while potentially reducing to a necessary minimum the need for valuable testing time with actual visual prosthesis carriers.
Keywords: computational modeling • vision and action • quality of life