This repository contains the code of the project for the course Deep Learning 02456. The authors are Frederik Lykke Andersen, Peter Nørgaard and Nils Hofmann. The project aims to segment parts of an image of car for the purpose of damage detection within a certain object. For that purpose, two convolutional neural networks were analyzed. As a final result the deeplabv3 network was used. In the jupyter file the user can take a look at the source code. Due to data security reasons the data could not be uploaded. Furthermore, the repository contains the U-Net under the Network directory and an Experiment directory were all python files are stored which were used in the beginning but were discarded in the later process of the project.