LocalNorm: Robust Image Classification through Dynamically Regularized Normalization
arXiv:1902.06550
Abstract
While modern convolutional neural networks achieve outstanding accuracy on many image classification tasks, they are, compared to humans, much more sensitive to image degradation. Here, we describe a variant of Batch Normalization, LocalNorm, that regularizes the normalization layer in the spirit of Dropout while dynamically adapting to the local image intensity and contrast at test-time. We show that the resulting deep neural networks are much more resistant to noise-induced image degradation, improving accuracy by up to three times, while achieving the same or slightly better accuracy on non-degraded classical benchmarks. In computational terms, LocalNorm adds negligible training cost and little or no cost at inference time, and can be applied to already-trained networks in a straightforward manner.
14 pages, 17 figures
References in corpus (7)
- Very Deep Convolutional Networks for Large-Scale Image Recognition
- Instance Normalization: The Missing Ingredient for Fast Stylization
- Accurate, Large Minibatch SGD: Training ImageNet in 1 Hour
- Benchmarking Neural Network Robustness to Common Corruptions and Perturbations
- Differentiable Learning-to-Normalize via Switchable Normalization
- Three Mechanisms of Weight Decay Regularization
- Dynamically Unfolding Recurrent Restorer: A Moving Endpoint Control Method for Image Restoration