Distinguishing artefacts: evaluating the saturation point of convolutional neural networks

Open in new window