A Tutor For The Hearing Impaired (Developed Using Automatic Gesture Recognition)
Ila Agarwal, Swati Johar, Jayashree Santhosh · International Journal of Computer Science Engineering and Applications · 2011
Automatic gesture recognition could be used to provide feedback in a computer-based learning environment to practice sign language skills for the deaf.In the present work the gestures made by the deaf as a communication aid, as per the Indian Sign Language were captured and a method was evolved to recognize them as English alphabets.After segmenting and removing the noise, the gestures were recognized in two stages.In the first stage, a common feature vector was calculated for all the alphabets, which captured the properties of overall shape of the gesture.If the gesture passed the first stage, then in the second stage a feature vector was calculated for that part of the gesture that actually represented the alphabet.The approach was evaluated by training the module on data of five users and testing on data of another five users.The dataset was characterized by varied user characteristics including clothing and skin tones.The module is user-independent and could be used for education as well as for entertainment purposes.