This is a research I worked on in 2019/2020 at Politecnico of Tourin. In this work, we use an Incremental Learning Approach for learning one by one colorectal tissues' classes. We train a ResNet model both on all the classes together and also one by one, increasing the network size as a new class is taken into account. Finally, we compare the two sets of results and assess then it is possible to learn a set of limited histological classes in an incremental fashion without any significant drop in performance. Here's an extract of the final results: