Feature Binding with Category-Dependant MixUp for Semantic Segmentation and Adversarial Robustness

Abstract

In this paper, we present a strategy for training convolutional neural networks to effectively resolve interference arising from competing hypotheses relating to inter-categorical information throughout the network. The premise is based on the notion of feature binding, which is defined as the process by which activation's spread across space and layers in the network are successfully integrated to arrive at a correct inference decision. In our work, this is accomplished for the task of dense image labelling by blending images based on their class labels, and then training a feature binding network, which simultaneously segments and separates the blended images. Subsequent feature denoising to suppress noisy activations reveals additional desirable properties and high degrees of successful predictions. Through this process, we reveal a general mechanism, distinct from any prior methods, for boosting the performance of the base segmentation network while simultaneously increasing robustness to adversarial attacks.

arXiv commentAccepted to BMVC 2020 (Oral)

References in corpus (10)

• Rethinking Atrous Convolution for Semantic Image Segmentation
• Improved Regularization of Convolutional Neural Networks with Cutout
• CutMix: Regularization Strategy to Train Strong Classifiers with Localizable Features
• Joint Optimization of Masks and Deep Recurrent Neural Networks for Monaural Source Separation
• Data Augmentation by Pairing Samples for Images Classification
• FMix: Enhancing Mixed Sample Data Augmentation
• Remix: Rebalanced Mixup
• Label Refinement Network for Coarse-to-Fine Semantic Segmentation
• Gated Feedback Refinement Network for Coarse-to-Fine Dense Semantic Image Labeling
• Semantics Meet Saliency: Exploring Domain Affinity and Models for Dual-Task Prediction