Web based prototype skin cancer classification using CNN models that classifiying 7 class of skin cancer based on lesions skin condition.
Made by : Ali Rohman (H1A018021) Teknik Elektro Universitas Jenderal Soedirman
This repository contains the machine learning models, I have also created flask web based UI for the classification prediction.
best_model_cutix.ipynb : iPythonNotebook File which classification of skin cancer based on the CNN model architecture that I made myself.
best_model_cutix.h5 : Weights are then saved to this file for directly used for UI purpose.
app.py : Flask based UI file which helps in prediction of the image by running the 'best_model_cutix.h5' file in the backend for making prediction by getting image by the user and predict the output.
index.html : For basic Interface of the webpage.
main.css : For styling the web interface.
main.js : To make websites more dynamic and interactive and to make file uploads more interactive on the web.
uploads : It contains the test image that uploaded byt users.
Link to download datasets : https://www.kaggle.com/kmader/skin-cancer-mnist-ham10000 , https://www.kaggle.com/discdiver/mnist1000-with-one-image-folder (contains all images in 1 folder only)
Web framework : Flask
Tensorflow
Matplotlib
Keras
Numpy
Pandas
Scikit-learn
Step 1 : Run the ‘best_model_cutix.ipynb’ file using Google Colaboratory/Jupyter Notebook/Visual Studio Code
Step 2 : At the final step of Training the model, save that model in the same folder in which the ‘app.py’ file.
Step 3 : Set the path folder of saved Model in app.py.
Step 4 : Now, run ‘app.py’ file to get the User Interface of Model Prediction (Classification)
Step 5 : Follow the localhost link to open the User Interface in Web Browser and start to classify skin lesion conditions to know what types of skin cancer from the images that uploaded by users.