An ML based project that predicts whether the uploaded image is spliced or authentic based off of 3 publicly available datasets. A voting/ensemble system is used to decide the final prediction.
Read results.txt for individual dataset results.
1. Run pip install -r requirements.txt in folder in a terminal
2. Run python3 databasesetup.py if a .db file doesnt exist
3. Run python3 server.py
4. In browser type localhost:0000
Notes:
-> Plots are all available in base folder
-> predictor.py used to generate the models is present too
Thank you!