File size: 18,060 Bytes
e8f2571 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 |
# GCNet
> [GCNet: Non-local Networks Meet Squeeze-Excitation Networks and Beyond](https://arxiv.org/abs/1904.11492)
<!-- [ALGORITHM] -->
## Abstract
The Non-Local Network (NLNet) presents a pioneering approach for capturing long-range dependencies, via aggregating query-specific global context to each query position. However, through a rigorous empirical analysis, we have found that the global contexts modeled by non-local network are almost the same for different query positions within an image. In this paper, we take advantage of this finding to create a simplified network based on a query-independent formulation, which maintains the accuracy of NLNet but with significantly less computation. We further observe that this simplified design shares similar structure with Squeeze-Excitation Network (SENet). Hence we unify them into a three-step general framework for global context modeling. Within the general framework, we design a better instantiation, called the global context (GC) block, which is lightweight and can effectively model the global context. The lightweight property allows us to apply it for multiple layers in a backbone network to construct a global context network (GCNet), which generally outperforms both simplified NLNet and SENet on major benchmarks for various recognition tasks.
<div align=center>
<img src="https://user-images.githubusercontent.com/40661020/143887418-c5d2c906-d6c9-415d-bce7-c5355c20a10a.png" height="300"/>
</div>
## Introduction
By [Yue Cao](http://yue-cao.me), [Jiarui Xu](http://jerryxu.net), [Stephen Lin](https://scholar.google.com/citations?user=c3PYmxUAAAAJ&hl=en), Fangyun Wei, [Han Hu](https://sites.google.com/site/hanhushomepage/).
We provide config files to reproduce the results in the paper for
["GCNet: Non-local Networks Meet Squeeze-Excitation Networks and Beyond"](https://arxiv.org/abs/1904.11492) on COCO object detection.
**GCNet** is initially described in [arxiv](https://arxiv.org/abs/1904.11492). Via absorbing advantages of Non-Local Networks (NLNet) and Squeeze-Excitation Networks (SENet), GCNet provides a simple, fast and effective approach for global context modeling, which generally outperforms both NLNet and SENet on major benchmarks for various recognition tasks.
## Results and Models
The results on COCO 2017val are shown in the below table.
| Backbone | Model | Context | Lr schd | Mem (GB) | Inf time (fps) | box AP | mask AP | Config | Download |
| :-------: | :---: | :------------: | :-----: | :------: | :------------: | :----: | :-----: | :-----------------------------------------------------: | :--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------: |
| R-50-FPN | Mask | GC(c3-c5, r16) | 1x | 5.0 | | 39.7 | 35.9 | [config](./mask-rcnn_r50-gcb-r16-c3-c5_fpn_1x_coco.py) | [model](https://download.openmmlab.com/mmdetection/v2.0/gcnet/mask_rcnn_r50_fpn_r16_gcb_c3-c5_1x_coco/mask_rcnn_r50_fpn_r16_gcb_c3-c5_1x_coco_20200515_211915-187da160.pth) \| [log](https://download.openmmlab.com/mmdetection/v2.0/gcnet/mask_rcnn_r50_fpn_r16_gcb_c3-c5_1x_coco/mask_rcnn_r50_fpn_r16_gcb_c3-c5_1x_coco_20200515_211915.log.json) |
| R-50-FPN | Mask | GC(c3-c5, r4) | 1x | 5.1 | 15.0 | 39.9 | 36.0 | [config](./mask-rcnn_r50-gcb-r4-c3-c5_fpn_1x_coco.py) | [model](https://download.openmmlab.com/mmdetection/v2.0/gcnet/mask_rcnn_r50_fpn_r4_gcb_c3-c5_1x_coco/mask_rcnn_r50_fpn_r4_gcb_c3-c5_1x_coco_20200204-17235656.pth) \| [log](https://download.openmmlab.com/mmdetection/v2.0/gcnet/mask_rcnn_r50_fpn_r4_gcb_c3-c5_1x_coco/mask_rcnn_r50_fpn_r4_gcb_c3-c5_1x_coco_20200204_024626.log.json) |
| R-101-FPN | Mask | GC(c3-c5, r16) | 1x | 7.6 | 11.4 | 41.3 | 37.2 | [config](./mask-rcnn_r101-gcb-r16-c3-c5_fpn_1x_coco.py) | [model](https://download.openmmlab.com/mmdetection/v2.0/gcnet/mask_rcnn_r101_fpn_r16_gcb_c3-c5_1x_coco/mask_rcnn_r101_fpn_r16_gcb_c3-c5_1x_coco_20200205-e58ae947.pth) \| [log](https://download.openmmlab.com/mmdetection/v2.0/gcnet/mask_rcnn_r101_fpn_r16_gcb_c3-c5_1x_coco/mask_rcnn_r101_fpn_r16_gcb_c3-c5_1x_coco_20200205_192835.log.json) |
| R-101-FPN | Mask | GC(c3-c5, r4) | 1x | 7.8 | 11.6 | 42.2 | 37.8 | [config](./mask-rcnn_r101-gcb-r4-c3-c5_fpn_1x_coco.py) | [model](https://download.openmmlab.com/mmdetection/v2.0/gcnet/mask_rcnn_r101_fpn_r4_gcb_c3-c5_1x_coco/mask_rcnn_r101_fpn_r4_gcb_c3-c5_1x_coco_20200206-af22dc9d.pth) \| [log](https://download.openmmlab.com/mmdetection/v2.0/gcnet/mask_rcnn_r101_fpn_r4_gcb_c3-c5_1x_coco/mask_rcnn_r101_fpn_r4_gcb_c3-c5_1x_coco_20200206_112128.log.json) |
| Backbone | Model | Context | Lr schd | Mem (GB) | Inf time (fps) | box AP | mask AP | Config | Download |
| :-------: | :--------------: | :------------: | :-----: | :------: | :------------: | :----: | :-----: | :--------------------------------------------------------------------------------------: | :------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------: |
| R-50-FPN | Mask | - | 1x | 4.4 | 16.6 | 38.4 | 34.6 | [config](./mask-rcnn_r50-syncbn_fpn_1x_coco.py) | [model](https://download.openmmlab.com/mmdetection/v2.0/gcnet/mask_rcnn_r50_fpn_syncbn-backbone_1x_coco/mask_rcnn_r50_fpn_syncbn-backbone_1x_coco_20200202-bb3eb55c.pth) \| [log](https://download.openmmlab.com/mmdetection/v2.0/gcnet/mask_rcnn_r50_fpn_syncbn-backbone_1x_coco/mask_rcnn_r50_fpn_syncbn-backbone_1x_coco_20200202_214122.log.json) |
| R-50-FPN | Mask | GC(c3-c5, r16) | 1x | 5.0 | 15.5 | 40.4 | 36.2 | [config](./mask-rcnn_r50-syncbn-gcb-r16-c3-c5_fpn_1x_coco.py) | [model](https://download.openmmlab.com/mmdetection/v2.0/gcnet/mask_rcnn_r50_fpn_syncbn-backbone_r16_gcb_c3-c5_1x_coco/mask_rcnn_r50_fpn_syncbn-backbone_r16_gcb_c3-c5_1x_coco_20200202-587b99aa.pth) \| [log](https://download.openmmlab.com/mmdetection/v2.0/gcnet/mask_rcnn_r50_fpn_syncbn-backbone_r16_gcb_c3-c5_1x_coco/mask_rcnn_r50_fpn_syncbn-backbone_r16_gcb_c3-c5_1x_coco_20200202_174907.log.json) |
| R-50-FPN | Mask | GC(c3-c5, r4) | 1x | 5.1 | 15.1 | 40.7 | 36.5 | [config](./mask-rcnn_r50-syncbn-gcb-r4-c3-c5_fpn_1x_coco.py) | [model](https://download.openmmlab.com/mmdetection/v2.0/gcnet/mask_rcnn_r50_fpn_syncbn-backbone_r4_gcb_c3-c5_1x_coco/mask_rcnn_r50_fpn_syncbn-backbone_r4_gcb_c3-c5_1x_coco_20200202-50b90e5c.pth) \| [log](https://download.openmmlab.com/mmdetection/v2.0/gcnet/mask_rcnn_r50_fpn_syncbn-backbone_r4_gcb_c3-c5_1x_coco/mask_rcnn_r50_fpn_syncbn-backbone_r4_gcb_c3-c5_1x_coco_20200202_085547.log.json) |
| R-101-FPN | Mask | - | 1x | 6.4 | 13.3 | 40.5 | 36.3 | [config](./mask-rcnn_r101-syncbn_fpn_1x_coco.py) | [model](https://download.openmmlab.com/mmdetection/v2.0/gcnet/mask_rcnn_r101_fpn_syncbn-backbone_1x_coco/mask_rcnn_r101_fpn_syncbn-backbone_1x_coco_20200210-81658c8a.pth) \| [log](https://download.openmmlab.com/mmdetection/v2.0/gcnet/mask_rcnn_r101_fpn_syncbn-backbone_1x_coco/mask_rcnn_r101_fpn_syncbn-backbone_1x_coco_20200210_220422.log.json) |
| R-101-FPN | Mask | GC(c3-c5, r16) | 1x | 7.6 | 12.0 | 42.2 | 37.8 | [config](./mask-rcnn_r101-syncbn-gcb-r16-c3-c5_fpn_1x_coco.py) | [model](https://download.openmmlab.com/mmdetection/v2.0/gcnet/mask_rcnn_r101_fpn_syncbn-backbone_r16_gcb_c3-c5_1x_coco/mask_rcnn_r101_fpn_syncbn-backbone_r16_gcb_c3-c5_1x_coco_20200207-945e77ca.pth) \| [log](https://download.openmmlab.com/mmdetection/v2.0/gcnet/mask_rcnn_r101_fpn_syncbn-backbone_r16_gcb_c3-c5_1x_coco/mask_rcnn_r101_fpn_syncbn-backbone_r16_gcb_c3-c5_1x_coco_20200207_015330.log.json) |
| R-101-FPN | Mask | GC(c3-c5, r4) | 1x | 7.8 | 11.8 | 42.2 | 37.8 | [config](./mask-rcnn_r101-syncbn-gcb-r4-c3-c5_fpn_1x_coco.py) | [model](https://download.openmmlab.com/mmdetection/v2.0/gcnet/mask_rcnn_r101_fpn_syncbn-backbone_r4_gcb_c3-c5_1x_coco/mask_rcnn_r101_fpn_syncbn-backbone_r4_gcb_c3-c5_1x_coco_20200206-8407a3f0.pth) \| [log](https://download.openmmlab.com/mmdetection/v2.0/gcnet/mask_rcnn_r101_fpn_syncbn-backbone_r4_gcb_c3-c5_1x_coco/mask_rcnn_r101_fpn_syncbn-backbone_r4_gcb_c3-c5_1x_coco_20200206_142508.log.json) |
| X-101-FPN | Mask | - | 1x | 7.6 | 11.3 | 42.4 | 37.7 | [config](./mask-rcnn_x101-32x4d-syncbn_fpn_1x_coco.py) | [model](https://download.openmmlab.com/mmdetection/v2.0/gcnet/mask_rcnn_x101_32x4d_fpn_syncbn-backbone_1x_coco/mask_rcnn_x101_32x4d_fpn_syncbn-backbone_1x_coco_20200211-7584841c.pth) \| [log](https://download.openmmlab.com/mmdetection/v2.0/gcnet/mask_rcnn_x101_32x4d_fpn_syncbn-backbone_1x_coco/mask_rcnn_x101_32x4d_fpn_syncbn-backbone_1x_coco_20200211_054326.log.json) |
| X-101-FPN | Mask | GC(c3-c5, r16) | 1x | 8.8 | 9.8 | 43.5 | 38.6 | [config](./mask-rcnn_x101-32x4d-syncbn-gcb-r16-c3-c5_fpn_1x_coco.py) | [model](https://download.openmmlab.com/mmdetection/v2.0/gcnet/mask_rcnn_x101_32x4d_fpn_syncbn-backbone_r16_gcb_c3-c5_1x_coco/mask_rcnn_x101_32x4d_fpn_syncbn-backbone_r16_gcb_c3-c5_1x_coco_20200211-cbed3d2c.pth) \| [log](https://download.openmmlab.com/mmdetection/v2.0/gcnet/mask_rcnn_x101_32x4d_fpn_syncbn-backbone_r16_gcb_c3-c5_1x_coco/mask_rcnn_x101_32x4d_fpn_syncbn-backbone_r16_gcb_c3-c5_1x_coco_20200211_164715.log.json) |
| X-101-FPN | Mask | GC(c3-c5, r4) | 1x | 9.0 | 9.7 | 43.9 | 39.0 | [config](./mask-rcnn_x101-32x4d-syncbn-gcb-r4-c3-c5_fpn_1x_coco.py) | [model](https://download.openmmlab.com/mmdetection/v2.0/gcnet/mask_rcnn_x101_32x4d_fpn_syncbn-backbone_r4_gcb_c3-c5_1x_coco/mask_rcnn_x101_32x4d_fpn_syncbn-backbone_r4_gcb_c3-c5_1x_coco_20200212-68164964.pth) \| [log](https://download.openmmlab.com/mmdetection/v2.0/gcnet/mask_rcnn_x101_32x4d_fpn_syncbn-backbone_r4_gcb_c3-c5_1x_coco/mask_rcnn_x101_32x4d_fpn_syncbn-backbone_r4_gcb_c3-c5_1x_coco_20200212_070942.log.json) |
| X-101-FPN | Cascade Mask | - | 1x | 9.2 | 8.4 | 44.7 | 38.6 | [config](./cascade-mask-rcnn_x101-32x4d-syncbn_fpn_1x_coco.py) | [model](https://download.openmmlab.com/mmdetection/v2.0/gcnet/cascade_mask_rcnn_x101_32x4d_fpn_syncbn-backbone_1x_coco/cascade_mask_rcnn_x101_32x4d_fpn_syncbn-backbone_1x_coco_20200310-d5ad2a5e.pth) \| [log](https://download.openmmlab.com/mmdetection/v2.0/gcnet/cascade_mask_rcnn_x101_32x4d_fpn_syncbn-backbone_1x_coco/cascade_mask_rcnn_x101_32x4d_fpn_syncbn-backbone_1x_coco_20200310_115217.log.json) |
| X-101-FPN | Cascade Mask | GC(c3-c5, r16) | 1x | 10.3 | 7.7 | 46.2 | 39.7 | [config](./cascade-mask-rcnn_x101-32x4d-syncbn-r16-gcb-c3-c5_fpn_1x_coco.py) | [model](https://download.openmmlab.com/mmdetection/v2.0/gcnet/cascade_mask_rcnn_x101_32x4d_fpn_syncbn-backbone_r16_gcb_c3-c5_1x_coco/cascade_mask_rcnn_x101_32x4d_fpn_syncbn-backbone_r16_gcb_c3-c5_1x_coco_20200211-10bf2463.pth) \| [log](https://download.openmmlab.com/mmdetection/v2.0/gcnet/cascade_mask_rcnn_x101_32x4d_fpn_syncbn-backbone_r16_gcb_c3-c5_1x_coco/cascade_mask_rcnn_x101_32x4d_fpn_syncbn-backbone_r16_gcb_c3-c5_1x_coco_20200211_184154.log.json) |
| X-101-FPN | Cascade Mask | GC(c3-c5, r4) | 1x | 10.6 | | 46.4 | 40.1 | [config](./cascade-mask-rcnn_x101-32x4d-syncbn-r4-gcb-c3-c5_fpn_1x_coco.py) | [model](https://download.openmmlab.com/mmdetection/v2.0/gcnet/cascade_mask_rcnn_x101_32x4d_fpn_syncbn-backbone_r4_gcb_c3-c5_1x_coco/cascade_mask_rcnn_x101_32x4d_fpn_syncbn-backbone_r4_gcb_c3-c5_1x_coco_20200703_180653-ed035291.pth) \| [log](https://download.openmmlab.com/mmdetection/v2.0/gcnet/cascade_mask_rcnn_x101_32x4d_fpn_syncbn-backbone_r4_gcb_c3-c5_1x_coco/cascade_mask_rcnn_x101_32x4d_fpn_syncbn-backbone_r4_gcb_c3-c5_1x_coco_20200703_180653.log.json) |
| X-101-FPN | DCN Cascade Mask | - | 1x | | | 47.5 | 40.9 | [config](./cascade-mask-rcnn_x101-32x4d-syncbn-dconv-c3-c5_fpn_1x_coco.py) | [model](https://download.openmmlab.com/mmdetection/v2.0/gcnet/cascade_mask_rcnn_x101_32x4d_fpn_syncbn-backbone_dconv_c3-c5_1x_coco/cascade_mask_rcnn_x101_32x4d_fpn_syncbn-backbone_dconv_c3-c5_1x_coco_20210615_211019-abbc39ea.pth) \| [log](https://download.openmmlab.com/mmdetection/v2.0/gcnet/cascade_mask_rcnn_x101_32x4d_fpn_syncbn-backbone_dconv_c3-c5_1x_coco/cascade_mask_rcnn_x101_32x4d_fpn_syncbn-backbone_dconv_c3-c5_1x_coco_20210615_211019.log.json) |
| X-101-FPN | DCN Cascade Mask | GC(c3-c5, r16) | 1x | | | 48.0 | 41.3 | [config](./cascade-mask-rcnn_x101-32x4d-syncbn-dconv-c3-c5-r16-gcb-c3-c5_fpn_1x_coco.py) | [model](https://download.openmmlab.com/mmdetection/v2.0/gcnet/cascade_mask_rcnn_x101_32x4d_fpn_syncbn-backbone_dconv_c3-c5_r16_gcb_c3-c5_1x_coco/cascade_mask_rcnn_x101_32x4d_fpn_syncbn-backbone_dconv_c3-c5_r16_gcb_c3-c5_1x_coco_20210615_215648-44aa598a.pth) \| [log](https://download.openmmlab.com/mmdetection/v2.0/gcnet/cascade_mask_rcnn_x101_32x4d_fpn_syncbn-backbone_dconv_c3-c5_r16_gcb_c3-c5_1x_coco/cascade_mask_rcnn_x101_32x4d_fpn_syncbn-backbone_dconv_c3-c5_r16_gcb_c3-c5_1x_coco_20210615_215648.log.json) |
| X-101-FPN | DCN Cascade Mask | GC(c3-c5, r4) | 1x | | | 47.9 | 41.1 | [config](./cascade-mask-rcnn_x101-32x4d-syncbn-dconv-c3-c5-r4-gcb-c3-c5_fpn_1x_coco.py) | [model](https://download.openmmlab.com/mmdetection/v2.0/gcnet/cascade_mask_rcnn_x101_32x4d_fpn_syncbn-backbone_dconv_c3-c5_r4_gcb_c3-c5_1x_coco/cascade_mask_rcnn_x101_32x4d_fpn_syncbn-backbone_dconv_c3-c5_r4_gcb_c3-c5_1x_coco_20210615_161851-720338ec.pth) \| [log](https://download.openmmlab.com/mmdetection/v2.0/gcnet/cascade_mask_rcnn_x101_32x4d_fpn_syncbn-backbone_dconv_c3-c5_r4_gcb_c3-c5_1x_coco/cascade_mask_rcnn_x101_32x4d_fpn_syncbn-backbone_dconv_c3-c5_r4_gcb_c3-c5_1x_coco_20210615_161851.log.json) |
**Notes:**
- The `SyncBN` is added in the backbone for all models in **Table 2**.
- `GC` denotes Global Context (GC) block is inserted after 1x1 conv of backbone.
- `DCN` denotes replace 3x3 conv with 3x3 Deformable Convolution in `c3-c5` stages of backbone.
- `r4` and `r16` denote ratio 4 and ratio 16 in GC block respectively.
## Citation
```latex
@article{cao2019GCNet,
title={GCNet: Non-local Networks Meet Squeeze-Excitation Networks and Beyond},
author={Cao, Yue and Xu, Jiarui and Lin, Stephen and Wei, Fangyun and Hu, Han},
journal={arXiv preprint arXiv:1904.11492},
year={2019}
}
```
|