TransMatcher

TransMatcher: Deep Image Matching Through Transformers for Generalizable Person Re-identification

This is the official PyTorch code for the TransMatcher proposed in our paper [1].

<img src="TransMatcher_Thumbnail.png" width=600>

For further details, please read our paper, and a poster here.

Usage

It is based on the QAConv 2.0 code, and the requirements and usage are quite similar. For a quick run, please try the demo.sh. Ignore the accuracy of this demo, since it is only for validating that everything is OK to run.

Performance

Performance (%) of TransMatcher under direct cross-dataset evaluation without transfer learning or domain adaptation:

<table align="center"> <tr align="center"> <td rowspan="2">Training Data</td> <td rowspan="2">Method</td> <td colspan="2">CUHK03-NP</td> <td colspan="2">Market-1501</td> <td colspan="2">MSMT17</td> </tr> <tr align="center"> <td>Rank-1</td> <td>mAP</td> <td>Rank-1</td> <td>mAP</td> <td>Rank-1</td> <td>mAP</td> </tr> <tr align="center"> <td rowspan="2">Market</td> <td>QAConv 2.0</td> <td><b>16.4</b></td> <td><b>15.7</b></td> <td>-</td> <td>-</td> <td><b>41.2</b></td> <td><b>15.0</b></td> </tr> <tr align="center"> <td>TransMatcher</td> <td>22.2</td> <td>21.4</td> <td>-</td> <td>-</td> <td>47.3</td> <td>18.4</td> </tr> <tr align="center"> <td rowspan="2">MSMT</td> <td>QAConv 2.0</td> <td><b>20.0</b></td> <td><b>19.2</b></td> <td><b>75.1</b></td> <td><b>46.7</b></td> <td>-</td> <td>-</td> </tr> <tr align="center"> <td>TransMatcher</td> <td>23.7</td> <td>22.5</td> <td>80.1</td> <td>52.0</td> <td>-</td> <td>-</td> </tr> <tr align="center"> <td rowspan="2">MSMT (all)</td> <td>QAConv 2.0</td> <td><b>27.2</b></td> <td><b>27.1</b></td> <td><b>80.6</b></td> <td><b>55.6</b></td> <td>-</td> <td>-</td> </tr> <tr align="center"> <td>TransMatcher</td> <td>31.9</td> <td>30.7</td> <td>82.6</td> <td>58.4</td> <td>-</td> <td>-</td> </tr> <tr align="center"> <td rowspan="2">RandPerson</td> <td>QAConv 2.0</td> <td><b>14.8</b></td> <td><b>13.4</b></td> <td><b>74.0</b></td> <td><b>43.8</b></td> <td><b>42.4</b></td> <td><b>14.4</b></td> </tr> <tr align="center"> <td>TransMatcher</td> <td>17.1</td> <td>16.0</td> <td>77.3</td> <td>49.1</td> <td>48.3</td> <td>17.7</td> </tr> </table>

Contacts

Shengcai Liao
Inception Institute of Artificial Intelligence (IIAI)
shengcai.liao@inceptioniai.org

Citation

[1] Shengcai Liao and Ling Shao, "TransMatcher: Deep Image Matching Through Transformers for Generalizable Person Re-identification." In Neural Information Processing Systems (NeurIPS), 2021.

@article{Liao-NeurIPS2021-TransMatcher,
  author    = {Shengcai Liao and Ling Shao},
  title     = {{TransMatcher: Deep Image Matching Through Transformers for Generalizable Person Re-identification}},
  booktitle = {Neural Information Processing Systems (NeurIPS)},  
  year={2021}
}