Identifying contaminants in astronomical images using convolutional neural networks

In this work, we propose to use convolutional neural networks to detect contaminants in astronomical images. Each contaminant is treated in a one vs all fashion. Once trained, our network is able to detect various contaminants such as cosmic rays, hot and bad pixel defaults, persistence effects, satellite trails or fringe patterns in images of various field properties. The convolutional neural network is performing semantic segmentation: it can output a probability map, assigning to each pixel its probability to belong to the contaminant or the background class. Training and testing data have been gathered from real or simulated data.