Enhanced multi-dataset transfer learning method for unsupervised person re-identification using co-training strategy

作     者：Xian, Yuqiao Hu, Haifeng 

作者机构：Sun Yat Sen Univ Sch Elect & Informat Technol Higher Educ Mega Ctr Waihuan East Rd Guangzhou Guangdong Peoples R China 

出 版 物：《IET COMPUTER VISION》 (IET电脑视觉)

年 卷 期：2018年第12卷第8期

页      面：1219-1227页

核心收录：

学科分类：0808[工学-电气工程] 08[工学] 0812[工学-计算机科学与技术（可授工学、理学学位）] 

基　　金：National Natural Science Foundation of China [61673402, 61273270, 60802069] Natural Science Foundation of Guangdong [2017A030311029, 2016B010109002, 2015B090912001, 2016B010123005, 2017B090909005] Science and Technology Program of Guangzhou [201704020180, 201604020024] Fundamental Research Funds for the Central Universities of China 

主　　题：neural nets unsupervised learning image recognition image representation iterative methods pattern clustering enhanced multidataset transfer learning method unsupervised person re-identification co-training strategy progressive unsupervised co-learning iterative training process transferred models multiple source datasets discriminative person representations single model large-scale benchmark datasets CNN models labelled source datasets multiple convolutional neural network models soft labels target dataset clustering 

摘      要：This study proposes progressive unsupervised co-learning for unsupervised person re-identification by introducing a co-training strategy in an iterative training process. The authors method adopts an iterative training process to improve transferred models by iterating among clustering, selection, exchange, and fine-tuning. To solve the problem of transferring representations learned from multiple source datasets, their method utilises multiple convolutional neural network (CNN) models trained on different labelled source datasets by feeding soft labels obtained by clustering on target dataset to each other. The enhanced model can learn more discriminative person representations than the single model trained on multiple datasets. Experimental results on two large-scale benchmark datasets (i.e. DukeMTMC-reID and Market-1501) demonstrate that their method can enhance transferred CNN models by using more source datasets and is competitive to the state-of-the-art methods.