TLFS23 - Tamil Language Finger Spelling Image Dataset

Tamil is an Indian Language of an remarkable antiquity that stands proudly as one of the oldest languages in the world. We have compiled a dataset for Tamil Language Finger Spelling which has finger representations for all 247 Tamil letters. TLFS23 is an elaborate dataset that comprises 248 classes of at least 1000 images in each class contributing to a massive 2,55,155 images in total. Each class represents Tamil characters finger spelled by participants of a diverse age group. Tamil has 12 Uyir ezuthukal (vowels), 18 Mei ezuthukal (consonants), 216 Uyir Mei ezuthukal (combination of consonant and vowels) and 1 Ayudha ezuthu (special character). This dataset has unique binary equivalent representations for all 247 characters of the Tamil language. The dataset could be used by scholars and researchers to develop vision based systems to assist the speech disabled.