Toward Inclusive Communication: Bit Equivalent Smart Gloves for Tamil Language Finger Spelling Translation

Abstract

While research in gesture recognition on popular and widespread languages like American Sign Language and Indian Sign Language has always aided the speech and hearing impaired in communicating better, the vernacular languages have not been brought to the spotlight to have similar technological aids developed. We present a Sensor-based Bit Equivalent Smart gloves for Tamil language finger spelling. Developing finger spelling for the Tamil language celebrates cultural diversity and inclusivity and aids in cultural preservation over the long term. The Tamil language has 247 characters which fall under two broad categories, which the two gloves on each hand detect as bit equivalent values based on finger movements. These finger movements are articulated into words or sentences and then displayed on an Android-based app to work as a translator. Flex sensors are used to recognize these bit-equivalent values. The left and the right glove communicate through an RF-based transceiver, and the right glove communicates to the Android app through Bluetooth. Overall, this is a low-powered wearable device with much focus on mobility and real-time usability. Experimental results indicate an average accuracy for all the characters as 91%, which was tested with 100 users. The results show that the device has appreciable accuracy and can be improved for day-to-day usage.