search

huanghoujing / beyond-part-models

PCB of paper: Beyond Part Models: Person Retrieval with Refined Part Pooling, using Pytorch
331 stars 81 forks source link

有关README #7

Closed luyang-NWPU closed 6 years ago

luyang-NWPU commented 6 years ago

在README里提到： | | Rank-1 (%) | mAP (%) | R.R. Rank-1 (%) | R.R. mAP (%) | | Market1501 (Paper)| 92.40 | 77.30 | - | - |

可是我看论文中说的是：“In this paper, we report mAP = 81.6%, 69.2%, 57.5% and Rank-1 = 93.8%, 83.3% and 63.7% for Market- 1501, Duke and CUHK03, respectively, setting new state of the art on the three datasets. All the results are achieved under the single-query mode without re-ranking. Re- ranking methods will further boost the performance espe- cially mAP. For example, when “PCB+RPP” is combined with the method in [44], mAP and Rank-1 accuracy on Market-1501 increases to 91.9% and 95.1%, respectively.”

为什么您提到的论文数据不一致呢？是我看错了么...

ZongweiZhou1 commented 6 years ago

@luyang-NWPU 代码中没有实现RPP部分，你说的是PCB+RPP的结果，看Table1