Update README.md
parent
ca27f37dd8
commit
83d93ba6d3
|
@ -1,4 +1,3 @@
|
|||
<<<<<<< HEAD
|
||||
# Re-ranking Person Re-identification with k-reciprocal Encoding
|
||||
================================================================
|
||||
|
||||
|
@ -69,7 +68,7 @@ The new training/testing protocol split for CUHK03 in our paper is in the "evalu
|
|||
matlab CUHK03_extract_feature.m
|
||||
```
|
||||
|
||||
3. Evaluation with our re-ranking method
|
||||
3. Evaluation
|
||||
```Shell
|
||||
# evaluation for Market-1501
|
||||
matlab Market_1501_evaluation.m
|
||||
|
@ -85,7 +84,7 @@ You can download our pre-trained IDE models and IDE features, and put them in th
|
|||
|
||||
- IDE features [[Baiduyun]](https://pan.baidu.com/s/1c1TtKcw) [[Google drive]](https://drive.google.com/open?id=0B7TOZKXmIjU3ODhaRm8yN2QzRHc)
|
||||
|
||||
Using the above IDE models and IDE features, you can reproduce the results with our re-ranking method as follows:
|
||||
Using the above IDE models and IDE features, you can reproduce the results as follows:
|
||||
|
||||
- Market-1501
|
||||
|
||||
|
@ -116,4 +115,3 @@ If you have any questions about this code, please do not hesitate to contact us.
|
|||
|
||||
=======
|
||||
# person-re-ranking
|
||||
>>>>>>> 8479ff10372e05534e4294c41347581dd73ec201
|
||||
|
|
Loading…
Reference in New Issue