TY - JOUR
T1 - Cross-domain person reidentification using domain adaptation ranking SVMs
AU - Ma, Andy J.
AU - Li, Jiawei
AU - YUEN, Pong Chi
AU - Li, Ping
N1 - This work was supported by the Hong Kong Research Grants Council-General Research Fund through the Hong Kong Baptist University under Grant 212313 and Grant 12202514. The work of A. J. Ma and P. Li was supported in part by the Office of Naval Research, Arlington, VA, USA, under Grant N00014-13-1-0764 and in part by the Air Force Office of Scientific Research, Arlington, VA, USA, under Grant FA9550-13-1-0137.
PY - 2015/5
Y1 - 2015/5
N2 - This paper addresses a new person reidentification problem without label information of persons under nonoverlapping target cameras. Given the matched (positive) and unmatched (negative) image pairs from source domain cameras, as well as unmatched (negative) and unlabeled image pairs from target domain cameras, we propose an adaptive ranking support vector machines (AdaRSVMs) method for reidentification under target domain cameras without person labels. To overcome the problems introduced due to the absence of matched (positive) image pairs in the target domain, we relax the discriminative constraint to a necessary condition only relying on the positive mean in the target domain. To estimate the target positive mean, we make use of all the available data from source and target domains as well as constraints in person reidentification. Inspired by adaptive learning methods, a new discriminative model with high confidence in target positive mean and low confidence in target negative image pairs is developed by refining the distance model learnt from the source domain. Experimental results show that the proposed AdaRSVM outperforms existing supervised or unsupervised, learning or non-learning reidentification methods without using label information in target cameras. Moreover, our method achieves better reidentification performance than existing domain adaptation methods derived under equal conditional probability assumption.
AB - This paper addresses a new person reidentification problem without label information of persons under nonoverlapping target cameras. Given the matched (positive) and unmatched (negative) image pairs from source domain cameras, as well as unmatched (negative) and unlabeled image pairs from target domain cameras, we propose an adaptive ranking support vector machines (AdaRSVMs) method for reidentification under target domain cameras without person labels. To overcome the problems introduced due to the absence of matched (positive) image pairs in the target domain, we relax the discriminative constraint to a necessary condition only relying on the positive mean in the target domain. To estimate the target positive mean, we make use of all the available data from source and target domains as well as constraints in person reidentification. Inspired by adaptive learning methods, a new discriminative model with high confidence in target positive mean and low confidence in target negative image pairs is developed by refining the distance model learnt from the source domain. Experimental results show that the proposed AdaRSVM outperforms existing supervised or unsupervised, learning or non-learning reidentification methods without using label information in target cameras. Moreover, our method achieves better reidentification performance than existing domain adaptation methods derived under equal conditional probability assumption.
KW - adaptive learning
KW - domain adaptation
KW - Person re-identification
KW - ranking SVMs
KW - target positive mean
UR - http://www.scopus.com/inward/record.url?scp=84925046292&partnerID=8YFLogxK
U2 - 10.1109/TIP.2015.2395715
DO - 10.1109/TIP.2015.2395715
M3 - Journal article
C2 - 25622315
AN - SCOPUS:84925046292
SN - 1057-7149
VL - 24
SP - 1599
EP - 1613
JO - IEEE Transactions on Image Processing
JF - IEEE Transactions on Image Processing
IS - 5
M1 - 7018030
ER -