Hi,I've been confused with the testing results on CUHK03 for a long time .WHY the accuracy of rank-1 is only around 30% ?It seems that the method does not work on the CUHK03 whether the CUHK03 is detected or not ?

layumi / Pedestrian_Alignment

TCSVT2018 Pedestrian Alignment Network for Large-scale Person Re-identification

https://ieeexplore.ieee.org/document/8481710

MIT License

236 stars 66 forks source link

Hi,I've been confused with the testing results on CUHK03 for a long time .WHY the accuracy of rank-1 is only around 30% ?It seems that the method does not work on the CUHK03 whether the CUHK03 is detected or not ? #11

Open niningxi opened 7 years ago

layumi commented 7 years ago

@niningxi Hi, we used the new protocol in CVPR2017 (https://github.com/zhunzhong07/person-re-ranking). The candidate gallery (5,332 images) is much larger than the former setting (100 images). So the result seems to be low.