Real-Time Detection and Reading of LED/LCD Displays for Visually Impaired Persons
By Tekin, Ender; Coughlan, James M.; Shen, Huiying; Proceedings of the 2011 IEEE Workshop on Applications of Computer Vision (WACV), pages 491-496Publication Date: January 5, 2011
Paper describes a novel system that acquires video, detects and reads LED/LCD characters in real time, reading them aloud to the user with synthesized speech. Liquid crystal displays (LCDs) or displays using light emitting diodes (LED), which are increasingly used in modern household appliances, are difficult or impossible to read by persons with blindness or visual impairment. Devices enabled by optical character recognition (OCR), while reading printed text, are not designed to read characters in appliance displays. The prototype display reader system reads 7-segment LED/LCD displays on (1) a desktop computer using a webcam running Windows XP, and (2) The Nokia N95 cell phone running Symbian OS. The computer vision algorithms underlying the system consist of a novel “blob” feature detection system which quickly extracts candidate digit features in the image, followed by a Hough-type voting scheme to classify each blob as a digit or non-digit, and then a grouping stage to determine the presence of coherent strings of LED/LCD display characters. The algorithms run at real time, achieving over 15 frames per second on the desktop, and around 5 frames per second on the Nokia N95. Experimental results are reported on videos of display images, demonstrating the feasibility of the system. Future directions for further development of the system are also discussed.
Published by:
Institute of Electrical & Electronics Engineers (IEEE) (Web Site: http://www.ieee.org )
Link to text: http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3146550/
