We are handairupa, your personal skincare assistant.
This repo contains all of our files and steps needed to replicate it.
We used skin90 from Kaggle as the dataset.
Model is created for image classification using transfer learning from InceptionV3.
Model is trained with k-fold cross validation, reaching 84% accuracy (our set threshold). Refer to the model_training.ipynb file inside the ML folder for complete documentation.
We used FastAPI for deployment.
Please refer to the handairupaApp for complete information.
TBA