Head mounted device for semantic representation of the user surroundings
A visually impaired person requires proper navigation as well as detailed description of the surroundings while traveling. Scientists through out the world wanted to develop some device to solve this problem. One such invention, describes a camera based object detection system for visually impaired persons where the user can identify the surrounding objects and independently navigate but he/she does not get a semantic or qualitative description of the surroundings. Scientists have also developed an apparatus to serve the same purpose where the visually impaired person can identify the adjacent objects but cannot get the perception of the surroundings while traversing.
Prof. S Chaudhuri and his students of Department of Electrical Engineering have worked to develop a camera based device system for visually impaired persons which would give a semantic description of the surroundings. It is economical and light weight wearable (head mounted) device for a visually impaired person and gives him/her a semantic representation of the surroundings traversed by him/her, i.e a verbal narration of the surroundings.This helps the user to form an image of the neighboring vicinity. Furthermore, this device can be expanded to a robot navigation system where the remote operator gets a perception of the surroundings traversed by the robot.
The device is comprised of a cap adapted to be fitted on the head of the user (if head mounted, else to be appropriately worn), an image capturing means mounted on the cap, a computing platform connected to the image capturing means, a feature data store attached to the computing platform and an output module connected to the computer module (Fig. 1).
The computing platform is comprised of a laptop PC. The images of the surroundings are captured either by a non-calibrated digital camera or an omni directional camera.
The feature data is stored on a USB drive or hard disk. The digital camera takes the photographs of the images of the surroundings of the user reflected on the outer surface of the cap which is coated with reflective material like mercuric oxide. The omni directional camera takes the images itself and those are divided into several sectors. The laptop PC extracts the color component of the images taken by the camera, compares it with the reference feature and identifies the corresponding neighboring object. A narrative description associated with the identified feature is fed either to the speakers or to a Braille board display. As a result the user is enabled to perceive the surroundings traversed by him. When this device was tested on a visually impaired person, the estimated description of the surroundings matched with the reality thereby establishing that the newly invented device is precise and dependable.