A machine learning based approach for the detection and recognition of Bangla sign language

Speech impaired people are detached from the mainstream society due to the lacking of proper communication aid. Sign language is the primary means of communication for them which normal people do not understand. In order to facilitate the conversation conversion of sign language to audio is very necessary. This paper aims at conversion of sign language to speech so that disabled people have their own voice to communicate with the general people. In this paper, Hand Gesture recognition is performed using HOG (Histogram of Oriented Gradients) for extraction of features from the gesture image and SVM (Support Vector Machine) as classifier. Finally, predict the gesture image with output text. This output text is converted into audible sound using TTS (Text to Speech) converter.