Flask App for Skin Cancer classification using IRV2 and Soft Attention

A flask app to run cancer classifier and classifiy cancer into 7 given classes:

• akiec: Actinic Keratoses and Intraepithelial Carcinoma/Bowen disease
• bcc: Basal Cell Carcinoma
• bkl: Benign lesions of the Keratosis type (solar lentigine/seborrheic keratoses and lichen-planus like keratosis)
• df: Dermatofibroma
• mel: Melanoma
• nv: Melanocytic Nevi
• vasc: Vascular Lesions (angiomas, angiokeratomas, pyogenic granulomas and hemorrhages)

File descriptions:

• The app.py contains the main logic for the flask app.
• The example_image.jpg is the exmaple image which is sent in encoded form in base64 format.
• example_request.ipynb shows how to use this flask app.
• requirements.txt contains the main requirements for the project.
• saved_model_v4.hdf5 is the actual model being used for the classification.

How to run:

• First install Python 3.11.6.
• Install everything mentioned in requirements.txt.
• Install any additional libs if any error occurs.
• Now, in the root directroy run the flask run command.
• Now, in the run the jupyter notebook to see the response as output.

This project was made more as microservice rather than being a standalone service so it might not meet your taste but it demonstrates the required functionality. You can see the example notebook for more info.

More to read

If you want to dive deeper into the model code and soft attention then have a look at this:

• paper 1
• paper 2 along with code here
• The dataset used in this model was a balanced subset of HAM10000 dataset.