aamnabhatti / cv_project-aamnabhatti-and-ameeraarif

Urdu is a complex language as it is an amalgam of many South Asian and East Asian languages; hence character recognition of Urdu is a huge and difficult task. We have systematically studied a recognition problem of handwritten Urdu digits. Our contribution is as follows:

1. Firstly we visualize our dataset using PCA, t-SNE anda combination of PCA+t-SNE
2. We apply 3 different architecture combinations namely:

• ResNet version 1 and version 2
• ResNet version 1 with Squeeze and Excitation block, ResNet version 2 with Squeeze and Exci-tation block
• ResNet version 1 with Convolutional Block Atten-tion Module, ResNet version 2 with ConvolutionalBlock Attention Module3)
  

3. Lastly we visualized feature maps to understand what islearned in them

Links to explore:

• Dataset
• Code
• CheckPoints
• Results

Result:

Attempt	Base Model	SE	CBAM
ResNetV1	97.76%	98.3%	98.68%
ResNetV2	98.1%	98.5%	99.2%

Confusion Matrix of ResNet V2 with CBAM