Bangla Handwritten Digit Recognition (Machine Learning Project)

#####################################

Description:

Handwritten digit recognition is a technique for automatically recognizing and detecting handwritten digital data via Machine Learning models. To meet the demand for paperless offices and to greatly improve work efficiency, a proper Bangla handwritten digit recognition system needs to be researched and implemented. It is not easy to recognize Bengali handwritten digits due to differences in shape, size and writing style. In this paper, we used TensorFlow’s CNN Algorithm- ‘Keras Sequential Model API’ to recognize the Bangla handwritten digit. We used 6 layers in modeling. We were able to get 99.1% accuracy in the 6th approach in recognizing Bangla handwritten digits.

Drive Path:

"/Project/Project_BHDR"

Screenshot:

Dataset -

Preprocessing -

Model Summary -

Confusion Matrix -