This repo contains the second project report and code for the Neural Networks course at Leiden University in the academic year 2019-2020. Before implementing a neural network which reads the time from a custom clock image data set, we familiarize ourselves with MLPs and CNNs using the Mnist digit and fashion image data. The report includes a detailed explanation of the MLP and CNN architecture and parameter choices to provide a theortical practical basis to create a network that is able to read the time. To read the time we compare 3 approchaes, a classifcation approach, a regression approach and a combination of both classifcation (for hours) and regression (minuets).