BengaliVQA : A Benchmark Dataset for Bengali Visual Question Answering
Published: 24 February 2025| Version 1 | DOI: 10.17632/y9fw6k37n9.1
Contributors:
, , Mahamudul Hasan Rafi, S.M. Hasan Imtiaz Labib, Faisal Muhammad ShahDescription
The Balanced Bengali VQA dataset consists of 3,280 images paired with 13,046 human-annotated yes/no questions and answers. The dataset is balanced to mitigate linguistic bias, offering a robust resource for Bengali visual question-answering research. This dataset serves as a valuable resource for advancing research in Bengali-language visual question answering.
Files
Institutions
- Bangladesh University of Engineering and Technology
- Ahsanullah University of Science and Technology
Categories
Computer Vision, Natural Language Processing, Object Detection