Person Re-id in the 3D Space

Thanks for your attention. In this repo, we provide the code for the paper [Person Re-identification in the 3D Space ].

News

• You may directly download my generated 3D data of the Market-1501 dataset at [OneDrive] or [GoogleDrive], and therefore you could skip the data preparation part.

Prerequisites

• Python 3.6 or 3.7
• GPU Memory >= 4G (e.g., GTX1080)
• Pytorch = 1.4.0 (Not Latest. Latest version is incompatible, since it changes the C++ interfaces.)
• dgl

Install

Here I use the cuda10.1 by default.

conda create --name OG python=3.7
conda activate OG
conda install pytorch=1.4.0 torchvision=0.5.0 cudatoolkit=10.1 -c pytorch
pip install dgl-cu101
pip install -r requirements.txt

If you face any error, you may first try to re-install open3d. It helps.

Prepare Data

• You may directly download my generated 3D data of the Market-1501 dataset at [OneDrive] or [GoogleDrive], and therefore you could skip the data preparation part.

Download Market-1501, DukeMTMC-reID or MSMT17 and unzip them in the ../

Split the dataset and arrange them in the folder of ID.

python prepare_market.py
python prepare_duke.py
python prepare_MSMT.py

Link the 2DDataset

ln -s ../Market/pytorch  ./2DMarket
ln -s ../Duke/pytorch  ./2DDuke
ln -s ../MSMT/pytorch  ./2DMSMT

Generate the 3D data via the code at https://github.com/layumi/hmr (I modified the code from https://github.com/akanazawa/hmr and added 2D-to-3D color mapping.)

Training

• Market-1501

OG-Net

python train_M.py --batch-size 8 --name ALL_Dense_b8_lr3.5_flip_slim0.5_warm5_scale_e0_d7+bg_adam_init768_clusterXYZRGB --slim 0.5 --flip --scale  --lrRate 3.5e-4 --gpu_ids 0 --warm_epoch 5  --erase 0  --droprate 0.7   --use_dense  --bg   --adam  --init 768  --cluster xyzrgb  --train_all

OG-Net-Small

python train_M.py --batch-size 8 --name ALL_SDense_b8_lr3.5_flip_slim0.5_warm5_scale_e0_d7+bg_adam_init768_clusterXYZRGB --slim 0.5 --flip --scale  --lrRate 3.5e-4 --gpu_ids 0 --warm_epoch 5  --erase 0  --droprate 0.7   --use_dense  --bg   --adam  --init 768  --cluster xyzrgb  --train_all     --feature_dims 48,96,192,384

• DukeMTMC-reID

OG-Net

python train_M.py --batch-size 8 --name ALL_Duke_Dense_b8_lr3.5_flip_slim0.5_warm5_scale_e0_d7+bg_adam_init768_clusterXYZRGB --slim 0.5 --flip --scale  --lrRate 3.5e-4 --gpu_ids 0 --warm_epoch 5  --erase 0  --droprate 0.7   --use_dense  --bg   --adam  --init 768  --cluster xyzrgb  --dataset-path 2DDuke  --train_all

OG-Net-Small

python train_M.py --batch-size 8 --name ALL_Duke_SDense_b8_lr3.5_flip_slim0.5_warm5_scale_e0_d7+bg_adam_init768_clusterXYZRGB --slim 0.5 --flip --scale  --lrRate 3.5e-4 --gpu_ids 0 --warm_epoch 5  --erase 0  --droprate 0.7   --use_dense  --bg   --adam  --init 768  --cluster xyzrgb  --train_all    --feature_dims 48,96,192,384 --dataset-path 2DDuke

• MSMT-17

OG-Net

python train_M.py --batch-size 8 --name MSMT_Dense_b8_lr3.5_flip_slim0.5_warm5_scale_e0_d7+bg_adam_init768_clusterXYZRGB --slim 0.5 --flip --scale  --lrRate 3.5e-4 --gpu_ids 0 --warm_epoch 5  --erase 0  --droprate 0.7   --use_dense  --bg   --adam  --init 768  --cluster xyzrgb  --dataset-path 2DMSMT

OG-Net-Small

python train_M.py --batch-size 8 --name ALL_MSMT_SDense_b8_lr3.5_flip_slim0.5_warm5_scale_e0_d5+bg_adam_init768_clusterXYZRGB --slim 0.5 --flip --scale  --lrRate 3.5e-4 --gpu_ids 0 --warm_epoch 5  --erase 0  --droprate 0.5   --use_dense  --bg   --adam  --init 768  --cluster xyzrgb  --dataset-path 2DMSMT  --train_all  --feature_dims 48,96,192,384

Evaluation

• Market-1501

python test_M.py  --name  ALL_Dense_b8_lr3.5_flip_slim0.5_warm5_scale_e0_d7+bg_adam_init768_clusterXYZRGB

• DukeMTMC-reID

python test_M.py  --data 2DDuke --name  ALL_Duke_Dense_b8_lr3.5_flip_slim0.5_warm5_scale_e0_d7+bg_adam_init768_clusterXYZRGB

• MSMT-17

python test_MSMT.py  --name MSMT_Dense_b8_lr3.5_flip_slim0.5_warm5_scale_e0_d7+bg_adam_init768_clusterXYZRGB

Pre-trained Models

Since OG-Net is really small, I has included trained models in this github repo ./snapshot.

Results

[Person Re-ID Performance]

Model name	Market	Duke	MSMT
OG-Net-Small	80.85(59.56)	70.11(49.93)	34.87(14.57)
OG-Net	80.94(59.97)	71.77(50.81)	36.37(15.74)

[ModelNet Performance]

I add OG-Net code to https://github.com/layumi/dgcnn
Results on ModelNet are 92.02 Top1 Accuracy / 88.84 MeanClass Top1 Accuracy.

Citation

You may cite it in your paper. Thanks a lot.

@article{zheng2020person,
  title={Person Re-identification in the 3D Space},
  author={Zhedong Zheng, Yi Yang},
  journal={arXiv 2006.04569},
  year={2020}
}

Related Work

We thank the great works of hmr, DGL, DGCNN and PointNet++. You may check their code at

• https://github.com/akanazawa/hmr
• https://github.com/dmlc/dgl/tree/master/examples/pytorch/pointcloud
• https://github.com/WangYueFt/dgcnn
• https://github.com/erikwijmans/Pointnet2_PyTorch

The baseline models used in the paper are modified from:

• https://github.com/layumi/Person_reID_baseline_pytorch

Acknowledge

I would like to thank the helpful comments and suggestions from Yaxiong Wang, Yuhang Ding, Qian Liu, Chuchu Han, Tianqi Tang, Zonghan Wu and Qipeng Guo.