Real-time sign language framework based on wearable device: analysis of MSL, DataGlove, and gesture recognition

Researchers have been inspired to use technology to enable people with hearing and speech impairment to communicate and engage with others around them. Sensory approach to recognition facilitates real-time and accurate recognition of signs. Thus, this study proposes a Malaysian Sign Language (MSL) recognition framework. The framework consists of three sub-modules for the recognition of static isolated signs based on data collected from a DataGlove. The first module focuses on the characteristics of signs, yielding sign recognition system requirements. The second module describes the different steps required to develop a wearable sign-capture device. The third module discusses the real-time SL recognition approach, which uses a template-matching algorithm to recognize acquired data. The final design of the DataGlove with 65 data channel fulfils the requirement identified from an analysis of MSL. The DataGlove is able to record data for all of the signs (both dynamic and static) of MSL due to the wide range of captured hand features. As a result, the recognition engine can accurately recognize complex signs.