Skip to content

Latest commit

 

History

History
77 lines (52 loc) · 10.1 KB

README.md

File metadata and controls

77 lines (52 loc) · 10.1 KB

Res2Net

Res2Net: A New Multi-scale Backbone Architecture

Abstract

Representing features at multiple scales is of great importance for numerous vision tasks. Recent advances in backbone convolutional neural networks (CNNs) continually demonstrate stronger multi-scale representation ability, leading to consistent performance gains on a wide range of applications. However, most existing methods represent the multi-scale features in a layer-wise manner. In this paper, we propose a novel building block for CNNs, namely Res2Net, by constructing hierarchical residual-like connections within one single residual block. The Res2Net represents multi-scale features at a granular level and increases the range of receptive fields for each network layer. The proposed Res2Net block can be plugged into the state-of-the-art backbone CNN models, e.g., ResNet, ResNeXt, and DLA. We evaluate the Res2Net block on all these models and demonstrate consistent performance gains over baseline models on widely-used datasets, e.g., CIFAR-100 and ImageNet. Further ablation studies and experimental results on representative computer vision tasks, i.e., object detection, class activation mapping, and salient object detection, further verify the superiority of the Res2Net over the state-of-the-art baseline methods.

Introduction

We propose a novel building block for CNNs, namely Res2Net, by constructing hierarchical residual-like connections within one single residual block. The Res2Net represents multi-scale features at a granular level and increases the range of receptive fields for each network layer.

Backbone Params. GFLOPs top-1 err. top-5 err.
ResNet-101 44.6 M 7.8 22.63 6.44
ResNeXt-101-64x4d 83.5M 15.5 20.40 -
HRNetV2p-W48 77.5M 16.1 20.70 5.50
Res2Net-101 45.2M 8.3 18.77 4.64

Compared with other backbone networks, Res2Net requires fewer parameters and FLOPs.

Note:

  • GFLOPs for classification are calculated with image size (224x224).

Results and Models

Faster R-CNN

Backbone Style Lr schd Mem (GB) Inf time (fps) box AP Config Download
R2-101-FPN pytorch 2x 7.4 - 43.0 config model | log

Mask R-CNN

Backbone Style Lr schd Mem (GB) Inf time (fps) box AP mask AP Config Download
R2-101-FPN pytorch 2x 7.9 - 43.6 38.7 config model | log

Cascade R-CNN

Backbone Style Lr schd Mem (GB) Inf time (fps) box AP Config Download
R2-101-FPN pytorch 20e 7.8 - 45.7 config model | log

Cascade Mask R-CNN

Backbone Style Lr schd Mem (GB) Inf time (fps) box AP mask AP Config Download
R2-101-FPN pytorch 20e 9.5 - 46.4 40.0 config model | log

Hybrid Task Cascade (HTC)

Backbone Style Lr schd Mem (GB) Inf time (fps) box AP mask AP Config Download
R2-101-FPN pytorch 20e - - 47.5 41.6 config model | log

Citation

@article{gao2019res2net,
  title={Res2Net: A New Multi-scale Backbone Architecture},
  author={Gao, Shang-Hua and Cheng, Ming-Ming and Zhao, Kai and Zhang, Xin-Yu and Yang, Ming-Hsuan and Torr, Philip},
  journal={IEEE TPAMI},
  year={2020},
  doi={10.1109/TPAMI.2019.2938758},
}