MGSR:MSL - An Integrated Mathematical GeometricalSpatial Reasoning with Mizo Sign Language forAugmented Cognitive and Communicative Synergies

Abstract

<title>Abstract</title> Sign language serves as a crucial means of communication for the deaf and mute community, facilitating their integration intosociety. There exist many sign languages in the world like ASL, ISL etc. in which much research have already been done. Thispaper represents the first effort to create a dataset for Mizo sign language, as no previous dataset exists for this low resourcelanguage. The work have been carried out by capturing hand gestures using high-quality camera Kinect V2. The differenthand gestures are captured on a different time periods such as morning, afternoon, evening and night time to capture therealistic scenarios faced by the deaf and dumb community. This dataset consists of 25 different sign alphabets with differenthand gestures for each alphabet in the Mizo languages. The original captured hand gestures in the dataset occupy a size of 3GB, the original data are resized to further enhance the computational time and better utilizes computational resources. Theresized data are then the fed through novel normalization technique to expand the dataset which results in a final dataset witha size of 10 GB. The uniqueness of our contribution lies in filling this gap by capturing and annotating sign language gesturesspecific to the Mizo language, which has not been documented in digital or structured formats before. This foundational step iscritical as it lays the groundwork for future research and development in sign language recognition for Mizo speakers.