hasanirtiza / PedesFormer-Transformer-Networks-For-Pedestrian-Detection

Transformer Networks for Pedestrian Detection

PedesFormer

PedesFormer is a MMDetection and SwinTransformer based repository. It is a successor to our earlier work Pedestron. PedesFormer, focuses on the adavancement of reseach on pedestrian detection using transformer networks.

🔥 Updates 🔥

• 🧨 Swin Transformer CityPerson model released. 🧨

Pretrained Models

Benchmarking

Benchmarking of pre-trained models on pedestrian detection datasets (autonomous driving)

Backbone	Dataset	Backbone	Configuration	Reasonable	Heavy
[Cascade Mask R-CNN]	CityPersons	Swin - Transformer	Config	9.2	36.9
[Cascade Mask R-CNN]	EuroCity Persons	Swin - Transformer	--
[Cascade Mask R-CNN]	Crowd Human	Swin - Transformer	--

More Pre-trained models are coming soon.

Installation

For installation, please see this.

Citation

Please cite the following works

CVPR2021

@InProceedings{Hasan_2021_CVPR,
    author    = {Hasan, Irtiza and Liao, Shengcai and Li, Jinpeng and Akram, Saad Ullah and Shao, Ling},
    title     = {Generalizable Pedestrian Detection: The Elephant in the Room},
    booktitle = {Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)},
    month     = {June},
    year      = {2021},
    pages     = {11328-11337}
}

ArXiv 2022

@article{hasan2022pedestrian,
  title={Pedestrian Detection: Domain Generalization, CNNs, Transformers and Beyond},
  author={Hasan, Irtiza and Liao, Shengcai and Li, Jinpeng and Akram, Saad Ullah and Shao, Ling},
  journal={arXiv preprint arXiv:2201.03176},
  year={2022}
}