Liming Zhao, Xi Li, Yueting Zhuang, and Jingdong Wang. “Deeply-Learned Part-Aligned Representations for Person Re-Identification.” Proceedings of the International Conference on Computer Vision (ICCV), 2017. (paper)
@InProceedings{Zhao_2017_ICCV,
author = {Zhao, Liming and Li, Xi and Zhuang, Yueting and Wang, Jingdong},
title = {Deeply-Learned Part-Aligned Representations for Person Re-Identification},
booktitle = {The IEEE International Conference on Computer Vision (ICCV)},
pages = {3219-3228},
year = {2017}
}
Contact: Liming Zhao (zlmzju@gmail.com)
-
Use my
Caffe
for using triplet loss layer. -
Run the demo code
demo/demo.ipynb
to see an example usage. -
Run
train.sh
in thetrain
folder to train the model. -
The datasets are placed in the
dataset
folder, you can download the archived data from here. Training list can be generated by using the code provided in the archieved data.
-
Use
Caffe
for implementation, please refer to the Caffe project website for installation. -
The protocal file in
proto
folder is written in python. -
The actual training scripts and protocal files will be generated in the
train
folder.