Ecosyste.ms: Awesome
An open API service indexing awesome lists of open source software.
https://github.com/hustvl/featurized-queryrcnn
Featurized Query R-CNN
https://github.com/hustvl/featurized-queryrcnn
Last synced: 3 days ago
JSON representation
Featurized Query R-CNN
- Host: GitHub
- URL: https://github.com/hustvl/featurized-queryrcnn
- Owner: hustvl
- License: mit
- Created: 2022-06-08T06:23:34.000Z (over 2 years ago)
- Default Branch: main
- Last Pushed: 2022-06-17T09:28:55.000Z (over 2 years ago)
- Last Synced: 2024-10-28T06:58:20.774Z (16 days ago)
- Language: Python
- Homepage: https://arxiv.org/abs/2206.06258
- Size: 1.26 MB
- Stars: 46
- Watchers: 5
- Forks: 3
- Open Issues: 3
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
Featurized Query R-CNN
Wenqiang Zhang*, Tianheng Cheng*, Xinggang Wang†, Shaoyu Chen, Qian Zhang, Wenyu Liu
(*: equal contribution, †: corresponding author)
## News
* `14 June, 2022`: Code and models of Featurized Query R-CNN have been released!
## Introduction
The query mechanism introduced in the DETR method is changing the paradigm of object detection and recently there are many query-based methods have obtained strong object detection performance. However, the current query-based detection pipelines suffer from the following two issues. Firstly, multi-stage decoders are required to optimize the randomly initialized object queries, incurring a large computation burden. Secondly, the queries are fixed after training, leading to unsatisfying generalization capability. To remedy the above issues, we present featurized object queries predicted by a query generation network in the well-established Faster R-CNN framework and develop a Featurized Query R-CNN. Extensive experiments on the COCO dataset show that our Featurized Query R-CNN obtains the best speed-accuracy trade-off among all R-CNN detectors, including the recent state-of-the-art Sparse R-CNN detector.
## Installation and Training
Our methods are based on [detectron2](), please refer to [here]() for more details.
Install the detectron2:
```
git clone https://github.com/facebookresearch/detectron2.git
python setup.py build develop
```
For training, run:
```
python train_net.py --config-file --num-gpus
```
## Main Results
| Model | Backbone | Epoch | AP | FPS | Weights |
| ---------------------------------------- | ---------- | ----- | ---- | ---- | ------- |
| [Featurized QR-CNN (100 Queries)](configs/queryrcnn.res50.100pro.3x.yaml) | ResNet-50 | 36 | 41.3 | 26 |[Google Drive](https://drive.google.com/file/d/19-UcZy8AtVghtjD8X9FrnftVqolJ0AwO/view?usp=sharing)|
| [Cascade Featurized QR-CNN (100 Queries)](configs/queryrcnn.cascade.res50.100pro.3x.yaml) | ResNet-50 | 36 | 43.0 | 24 |[Google Drive](https://drive.google.com/file/d/1bg_H4Ewaa8KkE31QLyhL8X6PnknRNYQT/view?usp=sharing)|
| [Cascade Featurized QR-CNN (300 Queries)](configs/queryrcnn.cascade.res50.300pro.3x.yaml) | ResNet-50 | 36 | 44.6 | 24 |[Google Drive](https://drive.google.com/file/d/1ZMH18xy7cfhPDYWL8kHoxMTWDLJXD5Y2/view?usp=sharing)|
| [Cascade Featurized QR-CNN (100 Queries)](configs/queryrcnn.cascade.res101.100pro.3x.yaml) | ResNet-101 | 36 | 43.9 | 18 |[Google Drive](https://drive.google.com/file/d/1oQFBhx5k8EeaDY34t_XhOMy8UrHrxj_U/view?usp=sharing)|
| [Cascade Featurized QR-CNN (300 Queries)](configs/queryrcnn.cascade.res101.300pro.3x.yaml) | ResNet-101 | 36 | 45.8 | 17 |[Google Drive](https://drive.google.com/file/d/1Epx1e7Xg9XQYsGmocBMlBHvZq5MtR4kr/view?usp=sharing)|
- The speed is tested on a single RTX 2080Ti GPU on COCO val set.
- If you have trouble accessing the models in Google Drive, we also provide the models in [BaiduPan](https://pan.baidu.com/s/1uLZI7iLtg9foZD8KzJ9hQw) for you (Password: **n91e**).
## Acknowledgements
Our implementation is based on [detectron2](https://github.com/facebookresearch/detectron2) and [Sparse R-CNN](), we thank for their open-source code.
## License
Featurized Query R-CNN is released under the [MIT Licence](LICENSE).
## Citation
If you find Featurized Query R-CNN is useful in your research or applications, please consider giving us a star 🌟 and citing it by the following BibTeX entry.
```bibtex
@article{FeaturizedQR-CNN,
title={Featurized Query R-CNN},
author={Zhang, Wenqiang and Cheng, Tianheng and Wang, Xinggang and Chen, Shaoyu and Zhang, Qian and Liu, Wenyu},
journal={arXiv preprint arXiv:2206.04584},
year={2022}
}
```