ChakmaLipī Dataset

Published: 13 March 2026| Version 6 | DOI: 10.17632/rzjphbvz9f.6
Contributors:
,
,
,
,
,

Description

This dataset contains 37,708 handwritten character images of the Chakma language, an endangered indigenous language of the Chittagong Hill Tracts in Bangladesh. The collection covers all 48 graphemes of the Chakma script, including 38 letters and 10 numerals. Specifications: Format: .png images Dimensions: 40 x 40 pixels Channels: 3 (RGB) Total Classes: 48 Total Samples: 37,708 Directory Structure: The root folder contains 48 sub-directories, each corresponding to a specific class label.

Files

Institutions

Categories

Computer Science, Computer Vision, Handwriting Recognition, Machine Learning

Licence