RobustNet: Improving Domain Generalization in Urban-Scene Segmentation via Instance Selective Whitening (2021-03-29T00:00:00.000000Z)

TL;DR

This paper proposes a novel instance selective whitening loss to improve the robustness of the segmentation networks for unseen domains and conducts extensive experiments in urban-scene segmentation to show the superiority of the approach to existing work.

Abstract

Enhancing the generalization capability of deep neural networks to unseen domains is crucial for safety-critical applications in the real world such as autonomous driving. To address this issue, this paper proposes a novel instance selective whitening loss to improve the robustness of the segmentation networks for unseen domains. Our approach disentangles the domain-specific style and domain-invariant content encoded in higher-order statistics (i.e., feature covariance) of the feature representations and selectively removes only the style information causing domain shift. As shown in Fig. 1, our method provides reasonable predictions for (a) low-illuminated, (b) rainy, and (c) unseen structures. These types of images are not included in the training dataset, where the baseline shows a significant performance drop, contrary to ours. Being simple yet effective, our approach improves the robustness of various backbone networks without additional computational cost. We conduct extensive experiments in urban-scene segmentation and show the superiority of our approach to existing work. Our code is available at this link1.

Authors

J. Choo

11 papers

Seungryong Kim

5 papers

Sungha Choi

1 papers

TL;DR

Abstract

Authors

References70 items

Whitening and second order optimization both destroy information about the dataset, and can make generalization impossible

Learning to Generate Novel Domains for Domain Generalization

Unsupervised Intra-Domain Adaptation for Semantic Segmentation Through Self-Supervision

Cars Can’t Fly Up in the Sky: Improving Urban-Scene Segmentation via Height-Driven Attention Networks

Correlation-aware Adversarial Domain Adaptation and Generalization

Domain Generalization via Model-Agnostic Learning of Semantic Features

Drop to Adapt: Learning Discriminative Features for Unsupervised Domain Adaptation

Domain Randomization and Pyramid Consistency: Simulation-to-Real Generalization Without Accessing Target Domain Data

Learning to Optimize Domain Specific Normalization for Domain Generalization

Depth-Attentional Features for Single-Image Rain Removal

Switchable Whitening for Deep Representation Learning

Iterative Normalization: Beyond Standardization Towards Efficient Whitening

Unsupervised Domain Adaptation Using Feature-Whitening and Consensus Loss

Feature-Critic Networks for Heterogeneous Domain Generalization

Episodic Training for Domain Generalization

Image-To-Image Translation via Group-Wise Deep Whitening-And-Coloring Transformation

DLOW: Domain Flow for Adaptation and Generalization

Improving Semantic Segmentation via Video Propagation and Label Relaxation

MetaReg: Towards Domain Generalization using Meta-Regularization

ADVENT: Adversarial Entropy Minimization for Domain Adaptation in Semantic Segmentation

Unsupervised Domain Adaptation for Semantic Segmentation via Class-Balanced Self-training

Deep Domain Generalization via Conditional Invariant Adversarial Networks

Diverse Image-to-Image Translation via Disentangled Representations

ShuffleNet V2: Practical Guidelines for Efficient CNN Architecture Design

Two at Once: Enhancing Learning and Generalization Capacities via IBN-Net

Effective Use of Synthetic Data for Urban Scene Semantic Segmentation

Whitening and Coloring Batch Transform for GANs

Domain Generalization with Adversarial Feature Learning

Exemplar Guided Unsupervised Image-to-Image Translation

BDD100K: A Diverse Driving Dataset for Heterogeneous Multitask Learning

Decorrelated Batch Normalization

DCAN: Dual Channel-wise Alignment Networks for Unsupervised Scene Adaptation

Multimodal Unsupervised Image-to-Image Translation

Learning to Adapt Structured Output Space for Semantic Segmentation

A Closed-form Solution to Photorealistic Image Stylization

Encoder-Decoder with Atrous Separable Convolution for Semantic Image Segmentation

MobileNetV2: Inverted Residuals and Linear Bottlenecks

Maximum Classifier Discrepancy for Unsupervised Domain Adaptation

Image to Image Translation for Domain Adaptation

ROAD: Reality Oriented Adaptation for Semantic Segmentation of Urban Scenes

CyCADA: Cycle-Consistent Adversarial Domain Adaptation

Learning to Generalize: Meta-Learning for Domain Generalization

The Mapillary Vistas Dataset for Semantic Understanding of Street Scenes

Unified Deep Supervised Domain Adaptation and Generalization

Learning Deep Architectures via Generalized Whitened Neural Networks

Curriculum Domain Adaptation for Semantic Segmentation of Urban Scenes

Rethinking Atrous Convolution for Semantic Image Segmentation

Universal Style Transfer via Feature Transforms

FCNs in the Wild: Pixel-level Adversarial and Constraint-based Adaptation

Pyramid Scene Parsing Network

Playing for Data: Ground Truth from Computer Games

Instance Normalization: The Missing Ingredient for Fast Stylization

Deep CORAL: Correlation Alignment for Deep Domain Adaptation

Image Style Transfer Using Convolutional Neural Networks

The SYNTHIA Dataset: A Large Collection of Synthetic Images for Semantic Segmentation of Urban Scenes

The Cityscapes Dataset for Semantic Urban Scene Understanding

Deep Residual Learning for Image Recognition

SegNet: A Deep Convolutional Encoder-Decoder Architecture for Image Segmentation

Domain Generalization for Object Recognition with Multi-task Autoencoders

ParseNet: Looking Wider to See Better

Domain-Adversarial Training of Neural Networks

Texture Synthesis Using Convolutional Neural Networks

U-Net: Convolutional Networks for Biomedical Image Segmentation

Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift

Fully convolutional networks for semantic segmentation

Unsupervised Domain Adaptation by Backpropagation

Domain Generalization via Invariant Feature Representation

Analysis of Representations for Domain Adaptation

Learning to Optimize Domain Speciﬁc Normalization for Domain Generalization

Medical Image Computing and Computer-Assisted Intervention

Field of Study

Journal Information

Name

Page

Venue Information

Name