This project has been conducted with my coworker Anas Hadhri during my last year at ESILV. The goal of this project is to build a predictive model which can classify painting styles. The model is trained on a database of 7500 pictures. All the data was retrieved from kaggle at this link : https://www.kaggle.com/paultimothymooney/collections-of-paintings-from-50-artists/data. Initially paintings were stored by painter, so a work around the dataset was performed in order to have paintings stored by style.

The dataset contains 12 styles which are among the most popular in the world. All paintings where the style was not clearly identify were drop from the dataset.

To train our model we used several techniques to compared the results : NN, CNN and Transfer learning.