Design Challenges in Effective Algorithm Development of Sign Language Recognition System

Sign language is the most putative language among hearing impaired people. They use non-verbal form of communication that involves hand gestures, head or body movement or facial expressions. Of these hand gestures is more widely used. Automatic Sign Language Recognition (ASLR) System can be used to convert these non-verbal signs into text or sound so that normal people can identify them without learning the sign language. ASLR employs Image Processing and Artificial Intelligence (AI) algorithms for effective conversion from sign to sound or text. This review unveils various image processing and AI steps involved in the conversion process, bringing out important topologies in the Image acquisition, segmentation, feature extraction, classification and detection process and a comparative analysis among various topologies. Attempts have been made to shed light into adoption of alternate design strategies in segmentation and feature extraction that enhance the performance in a complex environment.