A Push-Pull Layer Improves Robustness of Convolutional Neural Networks
read the original abstract
We propose a new layer in Convolutional Neural Networks (CNNs) to increase their robustness to several types of noise perturbations of the input images. We call this a push-pull layer and compute its response as the combination of two half-wave rectified convolutions, with kernels of opposite polarity. It is based on a biologically-motivated non-linear model of certain neurons in the visual system that exhibit a response suppression phenomenon, known as push-pull inhibition. We validate our method by substituting the first convolutional layer of the LeNet-5 and WideResNet architectures with our push-pull layer. We train the networks on nonperturbed training images from the MNIST, CIFAR-10 and CIFAR-100 data sets, and test on images perturbed by noise that is unseen by the training process. We demonstrate that our push-pull layers contribute to a considerable improvement in robustness of classification of images perturbed by noise, while maintaining state-of-the-art performance on the original image classification task.
This paper has not been read by Pith yet.
Forward citations
Cited by 1 Pith paper
-
Place recognition in gardens by learning visual representations: data set and benchmark analysis
An extended garden image dataset with pairwise ground truth is released and used to benchmark CNN place recognition models trained with contrastive loss, showing domain-specific improvement but limited generalization.
discussion (0)
Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.