Back to Volume
Paper: MaxiMask: Identifying Contaminants in Astronomical Images using Convolutional Neural Networks
Volume: 523, Astronomical Data Analysis Software and Systems XXVIII
Page: 99
Authors: Paillassa, M.; Bertin, E.; Bouy, H.
Abstract: We present MaxiMask, a contaminant detector for ground-based astronomical images based on convolutional neural networks (CNNs). Once trained, Maxi-Mask is able to detect cosmic rays, hot pixels, bad pixels, saturated pixels, diffraction spikes, nebulous features, persistence effects, satellite trails and residual fringe patterns in ground based images, encompassing a broad range of ambient conditions, PSF sampling, detectors, optics and stellar density. Individual image pixels can be flagged through semantic segmentation, based on high-resolution probability maps generated by MaxiMask for each contaminant, except for the tracking error probability which is assigned by another dedicated CNN. Training and testing data have been gathered from a large dataset of simulated and real data originating from various modern CCD and near-IR cameras.
Back to Volume