A precortical module for robust CNNs to light variations
This addresses robustness issues in image classification for computer vision applications, but appears incremental as it adds a module to existing CNN architectures.
The authors tackled the problem of CNNs being sensitive to light and contrast variations by adding a precortical module inspired by mammalian vision. They validated this approach on MNIST, FashionMNIST, and SVHN datasets, obtaining significantly more robust CNNs.
We present a simple mathematical model for the mammalian low visual pathway, taking into account its key elements: retina, lateral geniculate nucleus (LGN), primary visual cortex (V1). The analogies between the cortical level of the visual system and the structure of popular CNNs, used in image classification tasks, suggests the introduction of an additional preliminary convolutional module inspired to precortical neuronal circuits to improve robustness with respect to global light intensity and contrast variations in the input images. We validate our hypothesis on the popular databases MNIST, FashionMNIST and SVHN, obtaining significantly more robust CNNs with respect to these variations, once such extra module is added.