Human Computer Interaction Based on Gestural Cues Recognition/Sign Language to Text Conversion

Human computer interaction is very wide-ranging and diverse field regarding research and design activity. This interaction between humans and computer systems can be done through various methods. Gesture recognition offers a natural and intuitive way for interaction. It is a natural and effective mean of communication and interaction for hearing-impaired people. Gestural cue is a category of non-verbal communication in which noticeable body actions transfer specific messages. This paper presents a gesture recognition system for the development of a Human Computer Interaction (HCI) using Leap Motion Sensor (LMS). LMS is a device proficient with tracking hand motions or gestures. The objective of this research is development of an HCI system that will convert sign language to text for hearing impaired people. Through hand or body gestures, the disabled can easily convey their message to the caregiver or robot. Sign language has been known for providing natural and intuitive way to interact with computers or machines and robots. We are employing three recognition techniques of Sign Language to Text Conversion (SLTC) to determine the performance of the model. Artificial Neural Network (ANN), Geometric Template Matching and Cross Correlation techniques were employed for static gesture recognition and the best results were acquired from geometric template matching.