TDL.zip_distance_person re-id_quietlyzts_top-push_行人重识别

Last updated: 2016/7/7 This package implements Top-push Distance Learning (TDL) algorithm and provides a demo for applying TDL on PRID2011 and ILIDS-VID datasets. To see the experimental results, please run "Demo.m". The features of PRID2011 and ILIDS-VID used in our paper are provided in "./data". If you would like to learn the metric by TDL using your own data, please call the function "TDL_train". % TDL_train train the metric by TDL with input training data. % Last updated: 2016/7/6 % % Usage: % [L,pca_para]=TDL_train(x,CountOfSampleEachClass,pars); % % Input: % x - Feature matrix of training set of size d x n, in which d is feature dimension and n is the number of samples. The samples are sorted by labels. % CountOfSampleEachClass - a row vector containing the number of samples of each class, corresponding to the feature matrix % pars.outdim - dimension of TDL output (default: 120) % pars.alpha - weight of the push term in the objective function (ranged from 0 to 1, default: 0.2) % pars.maxiter - max iteration time (default: 1000) % pars.dem_pca - dimension of PCA output (default: n-1, n is the number of samples) % % Output: % L - the learned metric M=L'*L % pca_para.dim - reduced dimension of PCA % pca_para.projection - projection matrix of PCA If you have any questions, please contact us via email: wuancong@mail2.sysu.edu.cn. Citation for using this code: Jinjie You, Ancong Wu, Xiang Li and Wei-Shi Zheng. Top-push Video-based Person Re-identification. In CVPR, 2016.