Scaling Wide Residual Networks for Panoptic Segmentation

Abstract

The Wide Residual Networks (Wide-ResNets), a shallow but wide model variant of the Residual Networks (ResNets) by stacking a small number of residual blocks with large channel sizes, have demonstrated outstanding performance on multiple dense prediction tasks. However, since proposed, the Wide-ResNet architecture has barely evolved over the years. In this work, we revisit its architecture design for the recent challenging panoptic segmentation task, which aims to unify semantic segmentation and instance segmentation. A baseline model is obtained by incorporating the simple and effective Squeeze-and-Excitation and Switchable Atrous Convolution to the Wide-ResNets. Its network capacity is further scaled up or down by adjusting the width (i.e., channel size) and depth (i.e., number of layers), resulting in a family of SWideRNets (short for Scaling Wide Residual Networks). We demonstrate that such a simple scaling scheme, coupled with grid search, identifies several SWideRNets that significantly advance state-of-the-art performance on panoptic segmentation datasets in both the fast model regime and strong model regime.

Update experimental results

References in corpus (19)

• Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift
• Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks
• MobileNets: Efficient Convolutional Neural Networks for Mobile Vision Applications
• Rethinking Atrous Convolution for Semantic Image Segmentation
• Res2Net: A New Multi-scale Backbone Architecture
• Accurate, Large Minibatch SGD: Training ImageNet in 1 Hour
• ParseNet: Looking Wider to See Better
• Fully Convolutional Networks for Semantic Segmentation
• AutoAugment: Learning Augmentation Policies from Data
• High-Resolution Representations for Labeling Pixels and Regions
• Multi-Task Learning Using Uncertainty to Weigh Losses for Scene Geometry and Semantics
• Self-training with Noisy Student improves ImageNet classification
• DeeperLab: Single-Shot Image Parser
• DetectoRS: Detecting Objects with Recursive Feature Pyramid and Switchable Atrous Convolution
• Micro-Batch Training with Batch-Channel Normalization and Weight Standardization
• Learning to Fuse Things and Stuff
• Understanding Convolution for Semantic Segmentation
• Auto-Panoptic: Cooperative Multi-Component Architecture Search for Panoptic Segmentation
• SpatialFlow: Bridging All Tasks for Panoptic Segmentation

Cited by in corpus (9)

• Masked-attention Mask Transformer for Universal Image Segmentation
• Backbones-Review: Feature Extraction Networks for Deep Learning and Deep Reinforcement Learning Approaches
• MGNet: Monocular Geometric Scene Understanding for Autonomous Driving
• A Comprehensive Review of Modern Object Segmentation Approaches
• DeepLab2: A TensorFlow Library for Deep Labeling
• RT-K-Net: Revisiting K-Net for Real-Time Panoptic Segmentation
• Locally Enhanced Self-Attention: Combining Self-Attention and Convolution as Local and Context Terms
• Panoptic Segmentation: A Review
• Stateless actor-critic for instance segmentation with high-level priors