Dilated Residual Networks

Abstract

Convolutional networks for image classification progressively reduce resolution until the image is represented by tiny feature maps in which the spatial structure of the scene is no longer discernible. Such loss of spatial acuity can limit image classification accuracy and complicate the transfer of the model to downstream applications that require detailed scene understanding. These problems can be alleviated by dilation, which increases the resolution of output feature maps without reducing the receptive field of individual neurons. We show that dilated residual networks (DRNs) outperform their non-dilated counterparts in image classification without increasing the model's depth or complexity. We then study gridding artifacts introduced by dilation, develop an approach to removing these artifacts (`degridding'), and show that this further increases the performance of DRNs. In addition, we show that the accuracy advantage of DRNs is further magnified in downstream applications such as object localization and semantic segmentation.

Published at the Conference on Computer Vision and Pattern Recognition (CVPR 2017)

References in corpus (8)

• Deep Residual Learning for Image Recognition
• Multi-Scale Context Aggregation by Dilated Convolutions
• Going Deeper with Convolutions
• The Cityscapes Dataset for Semantic Urban Scene Understanding
• Fully Convolutional Networks for Semantic Segmentation
• DeepLab: Semantic Image Segmentation with Deep Convolutional Nets, Atrous Convolution, and Fully Connected CRFs
• Learning Deconvolution Network for Semantic Segmentation
• Understanding Convolution for Semantic Segmentation

Cited by in corpus (61)

• High-Resolution Representations for Labeling Pixels and Regions
• CyCADA: Cycle-Consistent Adversarial Domain Adaptation
• Deformable Convolutional Networks
• Deep High-Resolution Representation Learning for Visual Recognition
• RedNet: Residual Encoder-Decoder Network for indoor RGB-D Semantic Segmentation
• Maximum Classifier Discrepancy for Unsupervised Domain Adaptation
• Context Encoding for Semantic Segmentation
• Light-Weight RefineNet for Real-Time Semantic Segmentation
• Domain Stylization: A Strong, Simple Baseline for Synthetic to Real Image Domain Adaptation
• Domain adaptation based self-correction model for COVID-19 infection segmentation in CT images
• GResNet: Graph Residual Network for Reviving Deep GNNs from Suspended Animation
• Guided Upsampling Network for Real-Time Semantic Segmentation
• Vortex Pooling: Improving Context Representation in Semantic Segmentation
• Unsupervised Cross-Modality Domain Adaptation of ConvNets for Biomedical Image Segmentations with Adversarial Loss
• Global Second-order Pooling Convolutional Networks
• A Weakly Supervised Adaptive DenseNet for Classifying Thoracic Diseases and Identifying Abnormalities
• Image to Image Translation for Domain Adaptation
• Pre-Training of Deep Bidirectional Protein Sequence Representations with Structural Information
• TextureGAN: Controlling Deep Image Synthesis with Texture Patches
• Attention-based Context Aggregation Network for Monocular Depth Estimation
• Significance-aware Information Bottleneck for Domain Adaptive Semantic Segmentation
• AUNet: Attention-guided dense-upsampling networks for breast mass segmentation in whole mammograms
• PlaneNet: Piece-wise Planar Reconstruction from a Single RGB Image
• Rethinking Spatially-Adaptive Normalization
• RegNet: Learning the Optimization of Direct Image-to-Image Pose Registration
• Accelerating Sparse Deep Neural Networks
• Dense Dilated UNet: Deep Learning for 3D Photoacoustic Tomography Image Reconstruction
• Crowd Counting with Density Adaption Networks
• Cost-Sensitive Active Learning for Intracranial Hemorrhage Detection
• Look Before You Leap: Learning Landmark Features for One-Stage Visual Grounding
• Floor-SP: Inverse CAD for Floorplans by Sequential Room-wise Shortest Path
• Parallel Grid Pooling for Data Augmentation
• Exploring Multi-Branch and High-Level Semantic Networks for Improving Pedestrian Detection
• Accurate Automatic Segmentation of Amygdala Subnuclei and Modeling of Uncertainty via Bayesian Fully Convolutional Neural Network
• IDD: A Dataset for Exploring Problems of Autonomous Navigation in Unconstrained Environments
• Learning Motion in Feature Space: Locally-Consistent Deformable Convolution Networks for Fine-Grained Action Detection
• Learning with Hierarchical Complement Objective
• Greenhouse Segmentation on High-Resolution Optical Satellite Imagery using Deep Learning Techniques
• PlaneSegNet: Fast and Robust Plane Estimation Using a Single-stage Instance Segmentation CNN
• Fusing Saliency Maps with Region Proposals for Unsupervised Object Localization
• Bipartite Conditional Random Fields for Panoptic Segmentation
• Omni-GAN: On the Secrets of cGANs and Beyond
• Multi${}^3$Net: Segmenting Flooded Buildings via Fusion of Multiresolution, Multisensor, and Multitemporal Satellite Imagery
• Inductive Visual Localisation: Factorised Training for Superior Generalisation
• Inter-BMV: Interpolation with Block Motion Vectors for Fast Semantic Segmentation on Video
• Optical Neural Networks
• Spatially-Adaptive Filter Units for Deep Neural Networks
• Context-Constrained Accurate Contour Extraction for Occlusion Edge Detection
• An Artificial Intelligence-Based System to Assess Nutrient Intake for Hospitalised Patients
• Multichannel Semantic Segmentation with Unsupervised Domain Adaptation
• TraceCaps: A Capsule-based Neural Network for Semantic Segmentation
• Dense Prediction with Attentive Feature Aggregation
• Unsupervised Domain Adaptation Network with Category-Centric Prototype Aligner for Biomedical Image Segmentation
• Residual Pyramid FCN for Robust Follicle Segmentation
• Holistic Decomposition Convolution for Effective Semantic Segmentation of 3D MR Images
• Efficient Fusion of Sparse and Complementary Convolutions
• Appearance-Based Gaze Estimation Using Dilated-Convolutions
• Split-Merge Pooling
• Multi-scale Convolution Aggregation and Stochastic Feature Reuse for DenseNets
• Joint Face Hallucination and Deblurring via Structure Generation and Detail Enhancement
• Seeking an Optimal Approach for Computer-Aided Pulmonary Embolism Detection