- Aurora Cerabolini
- Andrea Malinverno
The dataset contains concrete images having cracks and it is divided in two classes called Positive and Negative. Images having cracks belong to the class «Positive» and the images without cracks belong to the class «Negative». Each class contains 20000 images, so in total there are 40000 images.
We want to build a Convolutional Neural Network to correctly classify the images belonging to the Positive class and the images belonging to the Negative one.
Please read the report for more details!