Kishor Sarawadekar
Indian Institute of Technology, India
Posters-Accepted Abstracts: J Electr Electron Syst
After formulation of Braille, no other technology has enabled blind and visually impaired people to communicate effectively with the digital technology. So, they feel disconnected with today�s cyber-world. Data published by World Health Organization reveals that about 285 million people are visually impaired round the globe. Out of this, about 19 million children (below age 15) are estimated to be visually impaired. Braille displays, scanner or reader, etc., are the output devices which blind people use to get connected with computers. But, there are limitations while blind people have to enter data in a computer. Entire QWERTY keyboard like functionality is not feasible with 6- or 8- dot Braille. Human-computer interaction techniques can be applied to solve such issues. We are developing a simple, cost effective, portable and easy to use writing support system to address these issues. While developing this system, we have proposed a new dactylology with which blinds can interact with the computers. A new feature extraction technique called as reduced shape signature (RSS) is introduced. This method reduces number of feature sets and it is invariant to rotation, translation as well as scaling. To discriminate intra-class gestures differential angle and polygonal area is computed along with RSS. Prototype model of the proposed writing support system is developed and algorithm for gesture recognition is implemented on the same. Real time experiments are performed and it is observed that the symbols recognition rate of the proposed system is 97.53%. At this stage, blind users can edit contents in Microsoft words using this experimental setup. Efforts are going on to add dynamic gesture functionality in future.
Email: to.skishor@gmail.com
Journal of Electrical & Electronic Systems received 733 citations as per Google Scholar report