AsTel (Offline Assamese and Telugu Handwritten Characters)

Various datasets for character recognition have been collected, explored, and utilized for numerous OCR tasks. India, being a linguistically diverse country, has seen extensive OCR research, although some regional languages, such as Assamese and Telugu, have been less explored due to the unavailability of suitable datasets. To address this gap, the AsTel dataset has been developed, comprising Assamese and Telugu handwritten characters collected offline using a paint mobile application. The dataset includes vowels, consonants, and compound characters from these languages, with 415 unique character classes for Assamese and 625 for Telugu. The Assamese characters were collected in Assam, while the Telugu characters were collected in Andhra Pradesh. In the name AsTel datase, As refers to the Assamese language and Tel refers to the Telugu language.