MYsign: Malaysian sign language recognition system

Sign language is an important tool used by the impairment people as their communication tool. It utilizes hand articulation, face expression and body movement to convey message. However, there were scarce people in the society capable to communicate with these impairment people due to little knowledge about sign language. Hence, this creates barrier for them to do communication in their daily activities. In addition, although there exists a computerized sign language translator but it is very limited to American and British sign languages. This creates an opportunity to develop a Malaysian Sign Language recognition system. Henceforth, this project aims to develop a Malaysian Sign Language (MSL) recognition system namely “MYsign: Malaysian Sign Language Recognition System” that can recognize Malaysian Sign Language. This project was developed using the Deep Learning technology utilizing the Long-Short Term Memory (LSTM) algorithm to train the model for detection and recognition of the hand gesture. “Assalamualaikum”, “Waalaikumsalam”, “Terima Kasih”, “Sama-Sama”, and “Maaf” were five classes of basic communication in Malaysian Sign Language that will be covered for this MYsign system. A total of 200 real-time videos from all classes were used to train the developed system. Additionally, system testing was done to evaluate the precision of the MSL basic communication sign language detection and recognition. Based on the testing result achieved, it is proved that the developed MYsign system is capable to perform good recognition.