CBD2023:A Hypercomplex Bangla Handwriting Character Recognition Dataset for Hierarchical Class Expansion Using Deep Learning

Published: 25 September 2023| Version 2 | DOI: 10.17632/p8988t5cwg.2
jabed omor bappi


This dataset comprises approximately 80,000 Bangla character images, meticulously divided into training and test sets. Within this extensive collection, researchers will find a total of 580 distinct character classes, including numerical characters seamlessly integrated into the dataset. This rich and diverse dataset offers an invaluable resource for anyone embarking on research in the field of Bangla character recognition. Whether you are exploring novel machine learning algorithms, developing deep neural networks, or conducting comparative studies, this dataset provides a robust foundation for your research endeavors. Its carefully curated contents and clear division between training and testing data ensure that researchers can readily access and harness this data to advance the frontier of Bangla character recognition.



Image Analysis