Oral-Cancer-Deep-Learning-

CNN on Oral Cancer Holistic Images

1. Augmented the data with flip and rotate.

2. Segmented the data into normal and cancerous.

3. Added cnn and fc layer with images.

4. Created a [cnn-pool-cnn-pool-fc-softmax] architecture and got 67% accuracy (overfitting?)

5. Currently trying to accomodate the transfer learning by adapting to inception-v3 net weights with total data set as test data.

6. To augment Holistic Data , doing color augmentation and normalization(in progress)

7. Transfer learning the inceptionv3 results(in progress)

8. Doing patch based cnn and classification based on the "Patch-based Convolutional Neural Network for Whole Slide Tissue Image Classification" Paper(in progress).