Tensorflow 2 implementation for D-Caps: Diagnosis Capsules
The model contains the following libraries and frameworks
- Tensorflow 2.4
- Paper- "Dynamic Routing Between Capsules": Link
- Paper "DIAGNOSING COLORECTAL POLYPS IN THE WILD WITH CAPSULE NETWORKS": arXiV
Paper by Rodney LaLonde, Pujan Kandel, Concetto Spampinato, Michael B. Wallace, and Ulas Bagci. - Blog Tutorial: Link
- GitHub reference: Link
Note:
The implementation is not completed without the last pooling layer and the reconstruction part
Note:
The implementation is quite different
- I changed some naming styles
- I embedded the routing inside the ConvCapsuleLayer
- The routing is nearly the same with some little modifications
- I used different squash function and it's also inside the ConvCapsuleLayer
- If using Binary Classifications (2 classes) the shape of the last layer will be different because of the tf.squeeze()