TY - GEN
T1 - Person re-identification with joint-loss
AU - Liu, Junqi
AU - Jiang, Na
AU - Zhou, Zhong
AU - Xu, Yue
N1 - Publisher Copyright:
© 2017 IEEE.
PY - 2017/7/2
Y1 - 2017/7/2
N2 - Person re-identification is a technique that search the given target in the video surveillance network. This technique has been widely pplied to security and surveillance system, and also become a esearch hotspot in computer vision. Person re-identification has been challenging due to the large number of cameras in the network and ariation in camera angles, illumination, occlusion and poses. In this paper, we proposed a person re-id approach that can resist occlusions and variations based on a human pose guided convolution neural network framework with joint loss functions. We extract local features from body parts localized by landmarks, merge it with global features to learn the similarity metric. Identification loss and pose-constrained triplet loss function are jointly employed to train the model. Our approach outperforms most state-of-The-Art methods on three large-scale datasets, with an accuracy of 83.31%, 86.1% and 72.6% on Cuhk03, Market1501 and Duke MTMC-reID respectively.
AB - Person re-identification is a technique that search the given target in the video surveillance network. This technique has been widely pplied to security and surveillance system, and also become a esearch hotspot in computer vision. Person re-identification has been challenging due to the large number of cameras in the network and ariation in camera angles, illumination, occlusion and poses. In this paper, we proposed a person re-id approach that can resist occlusions and variations based on a human pose guided convolution neural network framework with joint loss functions. We extract local features from body parts localized by landmarks, merge it with global features to learn the similarity metric. Identification loss and pose-constrained triplet loss function are jointly employed to train the model. Our approach outperforms most state-of-The-Art methods on three large-scale datasets, with an accuracy of 83.31%, 86.1% and 72.6% on Cuhk03, Market1501 and Duke MTMC-reID respectively.
KW - Deep learning
KW - Joint loss
KW - Person re-identification
KW - Pose estimation
UR - https://www.scopus.com/pages/publications/85067074651
U2 - 10.1109/ICVRV.2017.00010
DO - 10.1109/ICVRV.2017.00010
M3 - 会议稿件
AN - SCOPUS:85067074651
T3 - Proceedings - 2017 International Conference on Virtual Reality and Visualization, ICVRV 2017
SP - 1
EP - 6
BT - Proceedings - 2017 International Conference on Virtual Reality and Visualization, ICVRV 2017
PB - Institute of Electrical and Electronics Engineers Inc.
T2 - 7th International Conference on Virtual Reality and Visualization, ICVRV 2017
Y2 - 21 October 2017 through 22 October 2017
ER -