Matlab code for our ICCV 2013 work "Person Re-identification by Salience Matching".
Created by Rui Zhao, June, 2014
##Installation
- Download VIPeR dataset, and put the subfolders (/cam_a and /cam_b) into directory .\dataset\viper\
- Download CUHK01 dataset, and put the dataset at directory ./dataset/campus
- If you run on Windows system, you need to replace the '/' with '' in following script files, set_paths.m, ./code/normdata.m, ./code/initialcontext_general.m, ./demo_salmatch_cuhk01.m
##Demo Currently, one demo is provided
- demo_salmatch_cuhk01.m : perform evaluation over CUHK01 dataset
##Remark
- Running the demo_salmatch_cuhk01.m is supposed to achieve 30.04% at rank-1 accuracy on CUHK01 dataset
- The training / testing partition is generated following the approach SDALF
- Parallel Toolbox can accellerate the computation, use matlabpool if necessary
- The demo will generate large amount of intermediate data in the cache directory (around 25.5GB for CUHK01 demo)
- This demo was tested on MATLAB (R2012b), both 64-bit Win7 and 64-bit Linux with Intel Xeon 3.33 GHz CPU.
##Citing our works Please kindly cite our work in your publications if it helps your research:
@inproceedings{zhao13person,
Author = {Zhao, Rui and Ouyang, Wanli and Wang, Xiaogang},
Title = { Person Re-identification by Salience Matching},
booktitle = {IEEE International Conference on Computer Vision (ICCV)},
year = {2013}
}
And our previous work published in CVPR 2013:
@inproceedings{zhao13unsupervised,
Author = {Zhao, Rui and Ouyang, Wanli and Wang, Xiaogang},
Title = { Unsupervised Salience Learning for Person Re-identification },
booktitle = {IEEE Conference on Computer Vision and Pattern Recognition (CVPR)},
year = {2013}
}