Mukh-Biboron
Published: 13 November 2023| Version 2 | DOI: 10.17632/4hgyg8p7kg.2
Contributors:
Ahmed Faiyaz, Noor Mairukh Khan Arnob, Md Mubtasim Fuad, Baivab DasDescription
Mukh-Biboron is a dataset containing Bangla textual descriptions of images of faces. This dataset has 40 appropriate Bangla Facial Attributes and their annotations for over 200,000 images. Mukh-Biboron also contains 10 diverse Bangla captions of more than 200,000 facial images. Mukh-Biboron is derived from the publicly available CelebA dataset.
Files
Steps to reproduce
The Bangla captions were generated using the script available at this GitHub repository: https://github.com/afaiyaz006/Mukh-Biboron-Caption-Generator
Institutions
University of Asia Pacific
Categories
Natural Language Processing