ISL2022: A novel dataset creation on Indian sign language

Standard datasets of Indian sign language (ISL) which are double handed and single handed gestures are developed for research which helps in automated sign language recognition system. The datasets wm provide the researchers an opportunity to develop and investigate the machine learning and deep learning algorithms which are used to recognize the gestures automatically which in turn benefits the speech and hearing impaired population. Authors of this paper have contributed a large amount of standard datasets for ISL. The datasets named ISL2022 has a collection of 117 ISL gestures which includes alphabets & digits(labelled), weeks, months, academic and bank related words, commonly used sentence level gestures which are collected from 85 signers who belong to different age groups and gender which is recorded with difference in background, costume, distance and light variations. A comparative analysis of the created datasets with the existing datasets is highlighted in this paper.