fast-reid/projects/PartialReID

DSR in FastReID

Deep Spatial Feature Reconstruction for Partial Person Re-identification

Lingxiao He, Xingyu Liao

[CVPR2018] [BibTeX]

Foreground-aware Pyramid Reconstruction for Alignment-free Occluded Person Re-identification

Lingxiao He, Xingyu Liao [ICCV2019] [BibTeX]

Installation

First install FastReID following the .. and then put Partial Datasets in directory datasets. The whole framework of FastReID-DSR is

< img src="https://firebasestorage.googleapis.com/v0/b/firescript-577a2.appspot.com/o/imgs%2Fapp%2FSherlockWorkspace%2F1nVTE3Sn5c.jpg?alt=media&token=e7e9fcfc-4fc1-49c8-bcf4-c007028fdd25" width="700px" />

and the detail you can refer to

Training and Evaluation

To train a model, run:

python3 projects/PartialReID/train_net.py --config-file <config.yaml>

For example, to train the re-id network with IBN-ResNet-50 Backbone one should execute:

CUDA_VISIBLE_DEVICES='0,1,2,3' python3 projects/PartialReID/train_net.py --config-file 'projects/PartialReID/configs/partial_market.yml'

Results

Method	PartialREID	OccludedREID	PartialiLIDS
	Rank@1 (mAP)	Rank@1 (mAP)	Rank@1 (mAP)
DSR (CVPR’18)	73.7(68.1)	72.8(62.8)	64.3(58.1)
FPR (ICCV'19)	81.0(76.6)	78.3(68.0)	68.1(61.8)
FastReID-DSR	82.7(76.8)	81.6(70.9)	73.1(79.8)

Citing DSR and Citing FPR

If you use DSR or FPR, please use the following BibTeX entry.

@inproceedings{he2018deep,
  title={Deep spatial feature reconstruction for partial person re-identification: Alignment-free approach},
  author={He, Lingxiao and Liang, Jian and Li, Haiqing and Sun, Zhenan},
  booktitle={IEEE Conference on Computer Vision and Pattern Recognition (CVPR)},
  year={2018}
}
@inproceedings{he2019foreground,
  title={Foreground-aware Pyramid Reconstruction for Alignment-free Occluded Person Re-identification},
  author={He, Lingxiao and Wang, Yinggang and Liu, Wu and Zhao, He and Sun, Zhenan and Feng, Jiashi},
  booktitle={IEEE International Conference on Computer Vision (ICCV)},
  year={2019}
}