Wider or Deeper: Revisiting the ResNet Model for Visual Recognition

Abstract

The trend towards increasingly deep neural networks has been driven by a general observation that increasing depth increases the performance of a network. Recently, however, evidence has been amassing that simply increasing depth may not be the best way to increase performance, particularly given other limitations. Investigations into deep residual networks have also suggested that they may not in fact be operating as a single deep network, but rather as an ensemble of many relatively shallow networks. We examine these issues, and in doing so arrive at a new interpretation of the unravelled view of deep residual networks which explains some of the behaviours that have been observed experimentally. As a result, we are able to derive a new, shallower, architecture of residual networks which significantly outperforms much deeper models such as ResNet-200 on the ImageNet classification dataset. We also show that this performance is transferable to other problem domains by developing a semantic segmentation approach which outperforms the state-of-the-art by a remarkable margin on datasets including PASCAL VOC, PASCAL Context, and Cityscapes. The architecture that we propose thus outperforms its comparators, including very deep ResNets, and yet is more efficient in memory use and sometimes also in training time. The code and models are available at https://github.com/itijyou/ademxapp

Code available at: https://github.com/itijyou/ademxapp

References in corpus (5)

• Very Deep Convolutional Networks for Large-Scale Image Recognition
• Learning Deconvolution Network for Semantic Segmentation
• Places: An Image Database for Deep Scene Understanding
• BoxSup: Exploiting Bounding Boxes to Supervise Convolutional Networks for Semantic Segmentation
• The Loss Surface of Residual Networks: Ensembles and the Role of Batch Normalization

Cited by in corpus (16)

• Semantic Instance Segmentation with a Discriminative Loss Function
• The Reversible Residual Network: Backpropagation Without Storing Activations
• Automatic Skin Lesion Analysis using Large-scale Dermoscopy Images and Deep Residual Networks
• Automatic Liver Lesion Detection using Cascaded Deep Residual Networks
• Stacked Deconvolutional Network for Semantic Segmentation
• MaskLab: Instance Segmentation by Refining Object Detection with Semantic and Direction Features
• Sharing Residual Units Through Collective Tensor Factorization in Deep Neural Networks
• Knowledge Adaptation for Efficient Semantic Segmentation
• Box-driven Class-wise Region Masking and Filling Rate Guided Loss for Weakly Supervised Semantic Segmentation
• Efficient Segmentation: Learning Downsampling Near Semantic Boundaries
• A Curriculum Domain Adaptation Approach to the Semantic Segmentation of Urban Scenes
• Decoders Matter for Semantic Segmentation: Data-Dependent Decoding Enables Flexible Feature Aggregation
• A Multiscale Patch Based Convolutional Network for Brain Tumor Segmentation
• Rethinking Convolutional Semantic Segmentation Learning
• How to make a pizza: Learning a compositional layer-based GAN model
• Exploring and Exploiting Diversity for Image Segmentation