1

Improving Semantic Segmentation via Decoupled Body and Edge Supervision

Existing semantic segmentation approaches either aim to improve the object's inner consistency by modeling the global context, or refine objects detail along their boundaries by multi-scale feature fusion. In this paper, a new paradigm for semantic …

Invertible Image Rescaling

High-resolution digital images are usually downscaled to fit various display screens or save the cost of storage and bandwidth, meanwhile the post-upscaling is adpoted to recover the original resolutions or the details in the zoom-in images. However, …

Boosted Histogram Transform for Regression

We propose a boosting algorithm for regression problems called boosted histogram transform for regression (BHTR) based on histogram transforms composed of random rotations, stretchings, and translations.

Implicit Euler Skip Connections: Enhancing Adversarial Robustness via Numerical Stability

We explore the relationship between the adversarial robustness and numerical stability. Furthermore, we propose IE-Skips, which is a modification of the vaniila skip connections for Residual Network Families inspired by the implicit Euler method and we also theoretically and exmperimently prove the advantages of our structure under adversarial attacks.

Maximum-and-Concatenation Networks

We propose a novel multi-layer DNN structure termed MCN, which can approximate some class of continuous functions arbitrarily well even with highly sparse connection. We prove that the global minima of an $l$-layer MCN may be outperformed, at least can be attained, by simply increasing the network depth. More importantly, MCN could be easily appended to any of the many existing DNN and the augmented DNN will share the same property of MCN.

Normalized Loss Functions for Deep Learning with Noisy Labels

Robust loss functions are essential for training accurate deep neural networks (DNNs) in the presence of noisy (incorrect) labels. It has been shown that the commonly used Cross Entropy (CE) loss is not robust to noisy labels. Whilst new loss …

PDO-eConvs: Partial Differential Operator Based Equivariant Convolutions

We employ PDOs as steerable filters to design a system equivariant to E(n). We use numerical schemes to discretize this system and obtain PDO-eConvs, which achieve a quadratic order equivariance approximation in the discrete domain.

Improving Adversarial Robustness Requires Revisiting Misclassified Examples

Deep neural networks (DNNs) are vulnerable to adversarial examples crafted by imperceptible perturbations. A range of defense techniques have been proposed to improve DNN robustness to adversarial examples, among which adversarial training has been …

Skip Connections Matter: On the Transferability of Adversarial Examples Generated with ResNets

Skip connections are an essential component of current state-of-the-art deep neural networks (DNNs) such as ResNet, WideResNet, DenseNet, and ResNeXt. Despite their huge success in building deeper and more powerful DNNs, we identify a surprising …

Spatial Pyramid Based Graph Reasoning for Semantic Segmentation

The convolution operation suffers from a limited receptive filed, while global modeling is fundamental to dense prediction tasks, such as semantic segmentation. In this paper, we apply graph convolution into the semantic segmentation task and propose …