CNN classifier

Marsich Gaia

Implementation of an image classifier based on convolutional neural networks:

• Folder data: stores the provided dataset. The two folders train and test that are inside collect the images respectively for the training phase and the testing phase. The label of an image is given by the folder in which it is contained

• Folder miscellaneous: created with the idea to store some useful code to do some tests. It just contains dim_images.ipynb, used to determine the type of output of the layers in the network

• Folder side_code: contains the files CNN.py and train_test.py, where I collected some methods that were useful to tackle the assignment

• File CVPR.ipynb: main file. The development of the project is here. It exploits the files in the side_code folder and uses the datasets provided in data

• File Report_GaiaMarsich.pdf: report on the project