Kaggle Competition & 2040 Midterm: Bengali.Al Handwritten Grapheme Classification

The midterm project aims to develop a deep learning model which can identify the three constituent elements in a handwritten Bengli character.

Links for the Blogs can be found at

• Initial Blog:https://medium.com/@wangzhi62948620/bengali-al-handwritten-graphemes-classification-initial-blog-post-f2d87c94968a
• Midway Blog:https://medium.com/@wangzhi62948620/bengali-al-handwritten-graphemes-classification-82cb684f58b5
• Final Blog：https://medium.com/@wangzhi62948620/bengali-al-handwritten-graphemes-classification-final-blog-bd703df7d97b

The link for the inference Kaggle Kernel:

• Kaggle entry:https://www.kaggle.com/cengc13/bengali-handwritten-grapheme-inference

Packages

• python version-Python 3.6.9
• Deep learning models-tensorflow==2.1.0 and keras==2.2.4
• Training utility functions-scikit-learn==0.22.1
• Data analysis-pandas==1.0.1
• Linear algebra-numpy==1.18.1
• Visualizaiton-matplotlib==3.1.3
• Image preprocessing-opencv==4.2.0 and pillow==5.1.0
• Image augmentation albumentations==0.4.3

References

• Starter EDA and Basic Model Kaggle notebook 1
• Mode codes related to EDA Kaggle notebook 2