Bangla Character Recognition
Bangla Character Recognition
Programming Language : Python
Framework: Tensorflow , Keras , Pandas , Matplotlib, OpenCV
Output Produced by The model:
Output Produced by The model:
Model:
Model:
Modified VGG-16
Modified VGG-16
Training and Testing dataset details:
Training and Testing dataset details:
Train image size : (90704, 28, 28, 1)
Test image size : (917, 28, 28, 1)
Train label size : (90704, 122)
Test label size : (917, 122)
Training and Testing Error :
Training and Testing Error :
Training Loss : 0.03627835959196091
Training Acc : 0.9907942414283752
Validation Loss : 0.2444198578596115
Validation Accuracy : 0.9552890062332153
Citation :
Citation :
Rabby A.S.A., Haque S., Islam M.S., Abujar S., Hossain S.A. (2019) Ekush: A Multipurpose and Multitype Comprehensive Database for Online Off-Line Bangla Handwritten Characters. In: Santosh K., Hegadi R. (eds) Recent Trends in Image Processing and Pattern Recognition. RTIP2R 2018. Communications in Computer and Information Science, vol 1037. Springer, Singapore