A Heuristic Decision Tree Algorithm for Filipino Sign Language (FSL) Translator Implemented Using Accelerometers with Bluetooth Connectivity to a Host PC

The Accelerometer-based Filipino Sign Language (FSL) Translator is an innovation to bridge the communication gap between signers and non-signers. The static acceleration-sensing feature of SCA2010 accelerometers, which are infixed in the fingertips of the glove, allowed the recognition of the gestures formed. Data inputs coming from the accelerometers were passed to the ATMega2560 and compared to a gesture library, i.e. Decision Tree classifier. The Decision Tree algorithm is based on the data performed in 30 trials and is implemented to improve the speed recognition of the system. The design is capable of rendering fingerspelled alphabets into a readable text form displayed on a host PC. The sign language translator system converts the 26 alphabets of the FSL alphabets, "Mahal Kita", and two user-defined Backspace and Clear gestures and yields 39.35 seconds average recognition speed in performing all the gestures.