Are you sure you want to delete this task? Once this task is deleted, it cannot be recovered.
liuxinchen3 3bec0d4388 | 2 years ago | |
---|---|---|
configs | 2 years ago | |
datasets | 2 years ago | |
demo | 2 years ago | |
fastreid | 2 years ago | |
images | 2 years ago | |
projects | 2 years ago | |
tests | 2 years ago | |
tools | 2 years ago | |
.DS_Store | 2 years ago | |
.gitattributes | 2 years ago | |
GETTING_STARTED.md | 2 years ago | |
INSTALL.md | 2 years ago | |
LICENSE | 2 years ago | |
MODEL_ZOO.md | 2 years ago | |
README.md | 2 years ago |
SSKD is implemented based on FastReID v1.0.0, it provides a semi-supervised feature learning framework to learn domain-general representations. The framework is shown in
FastHuman is very challenging, as it contains more complex application scenarios and large-scale training, testing datasets. It has diverse images from different application scenarios including campus, airport, shopping mall, street, and railway station.
It contains 447,233 labeled images of 40,061 subjects captured by 82 cameras. The details of FastHuman, you can refer to paper.
Source Domain | #subjects | #images | #cameras | collection place |
---|---|---|---|---|
CUHK03 | 1,090 | 14,096 | 2 | campus |
SAIVT | 152 | 7,150 | 8 | buildings |
AirportALERT | 9,651 | 30,243 | 6 | airport |
iLIDS | 300 | 4,515 | 2 | airport |
PKU | 114 | 1,824 | 2 | campus |
PRAI | 1,580 | 39,481 | 2 | aerial imagery |
SenseReID | 1,718 | 3,338 | 2 | unknown |
SYSU | 510 | 30,071 | 4 | campus |
Thermalworld | 409 | 8,103 | 1 | unknown |
3DPeS | 193 | 1,012 | 1 | outdoor |
CAVIARa | 72 | 1,220 | 1 | shopping mall |
VIPeR | 632 | 1,264 | 2 | unknown |
Shinpuhkan | 24 | 4,501 | 8 | unknown |
WildTrack | 313 | 33,979 | 7 | outdoor |
cuhk-sysu | 11,934 | 34,574 | 1 | street |
LPW | 2,731 | 30,678 | 4 | street |
GRID | 1,025 | 1,275 | 8 | underground |
Total | 31,423 | 246,049 | 57 | - |
Unseen Domain | #subjects | #images | #cameras | collection place |
---|---|---|---|---|
Market1501 | 1,501 | 32,217 | 6 | campus |
DukeMTMC | 1,812 | 36,441 | 8 | campus |
MSMT17 | 4,101 | 126,441 | 15 | campus |
PartialREID | 60 | 600 | 6 | campus |
PartialiLIDS | 119 | 238 | 2 | airport |
OccludedREID | 200 | 2,000 | 5 | campus |
CrowdREID | 845 | 3,257 | 11 | railway station |
Total | 8,638 | 201,184 | 49 | - |
YouTube-Human is a unlabeled human dataset. You can download the Street-View video from YouTube website, and the use the human detection algorithm (centerX) to obtain the human images.
The whole training process is divided into two stages:
python3 projects/Basic_Project/train_net.py --config-file projects/Basic_Project/configs/r34-ibn.yml --num-gpu 4
python3 projects/Basic_Project/train_net.py --config-file projects/Basic_Project/configs/r101-ibn.yml --num-gpu 4
python3 projects/SSKD/train_net.py --config-file projects/SSKD/configs/sskd.yml --num-gpu 4
If you use fastreid or sskd in your research, please give credit to the following papers:
@article{he2020fastreid,
title={FastReID: A Pytorch Toolbox for General Instance Re-identification},
author={He, Lingxiao and Liao, Xingyu and Liu, Wu and Liu, Xinchen and Cheng, Peng and Mei, Tao},
journal={arXiv preprint arXiv:2006.02631},
year={2020}
}
@article{he2021semi,
title={Semi-Supervised Domain Generalizable Person Re-Identification},
author={He, Lingxiao and Liu, Wu and Liang, Jian and Zheng, Kecheng and Liao, Xingyu and Cheng, Peng and Mei, Tao},
journal={arXiv preprint arXiv:2108.05045},
year={2021}
}
该算法利用一种半监督知识蒸馏的方法来解决行人重识别的泛化问题,提出有标注数据和无标注数据的协同优化方法,提升模型整体的泛化能力。目前在多个数据库上接近于监督学习的性能。
C Python Protocol Buffer Markdown Cython
Dear OpenI User
Thank you for your continuous support to the Openl Qizhi Community AI Collaboration Platform. In order to protect your usage rights and ensure network security, we updated the Openl Qizhi Community AI Collaboration Platform Usage Agreement in January 2024. The updated agreement specifies that users are prohibited from using intranet penetration tools. After you click "Agree and continue", you can continue to use our services. Thank you for your cooperation and understanding.
For more agreement content, please refer to the《Openl Qizhi Community AI Collaboration Platform Usage Agreement》