two assumptions two assumptions 1) different video track-lets typically contain different persons。
2)within each tracklet, the frames are mostly of the same per-son. \ Our method is built onreciprocal nearest neighbor search and can eliminate thehard negative label matches, i.e., the cross-camera nearestneighbors of the false matches in the initial rank list. Thetracklet that passes the reciprocal nearest neighbor checkis considered to have the same ID with the query.
our workis motivated from three aspects. First, videos contain muchricher information than single images, Second, video tracklets, produced by pedestrian detec-tion and tracking (Fig. 1), are reliable data source for unsu-pervised learning methods. Third, feature learning using tracklets from the sameview may result in low discriminative ability.
In brief,two steps are involved:
1) classiﬁer initialization using thetracklets in the same camera
2) iterations between cross-camera tracklet association and featurelearning.
The proposed metric promotion approachiterates between model update and label estimation. Forlabel estimation, under camera A, we use each tracklet asquery to search for its k nearest neighbors (NNs) in cameraB. Among these k candidates, the best match is selected asbeing associated with the query tracklet. We employ neg-ative mining to reduce the impact of false positive matchesin the k-NNs. This k-NN search process is then reverselyrepeated using the best match as query to see whether theinitial query is its best match, a conﬁrmation protocol toensure that the initial query and the best match are truly as-sociated. The the associated pairs are adopted for modelupdating.
 Y. Cho and K. Yoon. Improving person re-identiﬁcation via pose-aware multi-shot matching. In CVPR, 2016.  J. Dai, Y. Zhang, and H. Lu. Cross-view semantic projection learningfor person re-identiﬁcation. In PR, 2017.  A. Dehghan, S. M. Assari, and M. Shah. GMMCP tracker: Globallyoptimal generalized maximum multi clique problem for multiple ob-ject tracking. In CVPR, 2015.  W. F and Z. C. Label propagation through linear neighborhoods. InTKDE, 2008.  H. Fan, L. Zheng, and Y. Yang. Unsupervised person re-identiﬁcation: Clustering and ﬁne-tuning. arXiv, 2017  M. Farenzena, L. Bazzani, A. Perina, V. Murino, and M. Cristani.Person re-identiﬁcation by symmetry-driven accumulation of localfeatures. In CVPR, 2010.  P. F. Felzenszwalb, R. B. Girshick, D. A. McAllester, and D. Ra-manan. Object detection with discriminatively trained part-basedmodels. TPAMI, 2010.  D. Gray and H. Tao. Viewpoint invariant pedestrian recognition withan ensemble of localized features. In ECCV, 2008.  O. Hamdoun, F. Moutarde, B. Stanciulescu, and B. Steux. Personre-identiﬁcation in multi-camera system by signature based on inter-est point descriptors collected on short video sequences. In ICDSC,2008.  J. F. Henriques, J. Carreira, R. Caseiro, and J. Batista. Beyond hardnegative mining: Efﬁcient detector learning via block-circulant de-composition. In ICCV, 2013.  M. Hirzer, C. Beleznai, P. M. Roth, and H. Bischof. Person re-identiﬁcation by descriptive and discriminative classiﬁcation. InSCIA, 2011.  S. Karanam, Y. Li, and R. J. Radke. Person re-identiﬁcation withdiscriminatively trained viewpoint invariant dictionaries. In ICCV,2015.  S. Karanam, Y. Li, and R. J. Radke. Sparse re-id: Block sparsity forperson re-identiﬁcation. In CVPR, 2015.  M. K ¨ostinger, M. Hirzer, P. Wohlhart, P. M. Roth, and H. Bischof.Large scale metric learning from equivalence constraints. In CVPR,2012.  N. Li, R. Jin, and Z. Zhou. Top rank optimization in linear time. InNIPS, 2014.  W. Li, R. Zhao, T. Xiao, and X. Wang. Deepreid: Deep ﬁlter pairingneural network for person re-identiﬁcation. In CVPR, 2014.  Y. Li, Z. Wu, S. Karanam, and R. J. Radke. Multi-shot human re-identiﬁcation using adaptive ﬁsher discriminant analysis. In BMVC,2015.  Z. Li, S. Chang, F. Liang, T. S. Huang, L. Cao, and J. R. Smith.Learning locally-adaptive decision functions for person veriﬁcation.In CVPR, 2013 . S. Liao, Y. Hu, X. Zhu, and S. Z. Li. Person re-identiﬁcation by localmaximal occurrence representation and metric learning. In CVPR,2015.  S. Liao, G. Zhao, V. Kellokumpu, M. Pietik¨ainen, and S. Z. Li. Mod-eling pixel process with scale invariant local patterns for backgroundsubtraction in complex scenes. In CVPR, 2010  C. Liu, C. C. Loy, S. Gong, and G. Wang. POP: person re-identiﬁcation post-rank optimisation. In ICCV, 2013.  K. Liu, B. Ma, W. Zhang, and R. Huang. A spatio-temporal appear-ance representation for viceo-based pedestrian re-identiﬁcation. InICCV, 2015. W. Liu and T. Zhang. Bidirectional label propagation over graphs.IJSI, 2013.  X. Liu, M. Song, D. Tao, X. Zhou, C. Chen, and J. Bu. Semi-supervised coupled dictionary learning for person re-identiﬁcation.In CVPR, 2014. B. M, H. S, and J. F. Hard negative mining for metric learning basedzero-shot classiﬁcation. In ECCV Workshops, 2016. B. Ma, Y. Su, and F. Jurie. Bicov: a novel image representation forperson re-identiﬁcation and face veriﬁcation. In BMVC, 2012.  B. Ma, Y. Su, and F. Jurie. Covariance descriptor based on bio-inspired features for person re-identiﬁcation and face veriﬁcation.IVC, 2014. X. Ma, X. Zhu, S. Gong, X. Xie, J. Hu, K.-M. Lam, and Y. Zhong.Person re-identiﬁcation by unsupervised video matching. PR, 2017. N. McLaughlin, J. Martinez del Rincon, and P. Miller. Recurrentconvolutional network for video-based person re-identiﬁcation. InCVPR, 2016.  M. Niall, M. del Rincon Jesus, and M. Paul. Recurrent convolutionalnetwork for video-based person re-identiﬁcation. In CVPR, 2016. S. Pedagadi, J. Orwell, S. A. Velastin, and B. A. Boghossian. Localﬁsher discriminant analysis for pedestrian re-identiﬁcation. In CVPR,2013.  B. J. Prosser, W. Zheng, S. Gong, and T. Xiang. Person re-identiﬁcation by support vector ranking. In BMVC, 2010.  P. Siva, C. Russell, and T. Xiang. In defence of negative mining forannotating weakly labeled data. In ECCV, 2012.  C. Sun, D. Wang, and H. Lu. Person re-identiﬁcation via distancemetric learning with latent variables. 2017.  M. Tetsu, O. Takahiro, S. Einoshin, and S. Yoichi. Hierarchical gaus-sian descriptor for person re-identiﬁcation. In CVPR, 2016 . R. Wang, S. Shan, X. Chen, and W. Gao. Manifold-manifold distancewith application to face recognition based on image set. In CVPR,2008.  T. Wang, S. Gong, X. Zhu, and S. Wang. Person re-identiﬁcation byvideo ranking. In ECCV, 2014.  T. Wang, S. Gong, X. Zhu, and S. Wang. Person re-identiﬁcation bydiscriminative selection in video ranking. TPAMI, 2016.  Y. Yan, B. Ni, Z. Song, C. Ma, Y. Yan, and X. Yang. Person re-identiﬁcation via recurrent feature aggregation. In ECCV, 2016 . M. Ye, C. Liang, Y. yu, and etal. Person re-identiﬁcation via rankingaggregation of similarity pulling and dissimilarity pushing. In TMM,2016.  M. Ye, J. Ma, J. Li, L. Zheng, and P. Yuen. Label graph matching forunsupervised video re-identiﬁcation. In ICCV, 2017.  J. You, A. Wu, X. Li, and W.-S. Zheng. Top-push video-based personre-identiﬁcation. In CVPR, 2016.  L. Zhang, T. Xiang, and S. Gong. Learning a discriminative nullspace for person re-identiﬁcation. In CVPR, 2016.  Y. Zhang, B. Li, and H. L. A. I. X. Ruan. Sample-speciﬁc svm learn-ing for person re-identiﬁcation. In CVPR, 2016 . Z. Zhang, X. Jing, and T. Wang. Label propagation based semi-supervised learning for software defect prediction. ASE, 2017.  Z. Zhang, M. Zhao, and T. W. S. Chow. Label propagation andsoft-similarity measure for graph based constrained semi-supervisedlearning. In IJCNN, 2014 . R. Zhao, W. Ouyang, and X. Wang. Unsupervised salience learningfor person re-identiﬁcation. In CVPR, 2013.  L. Zheng, Z. Bie, Y. Sun, J. Wang, C. Su, S. Wang, and Q. Tian.MARS: A video benchmark for large-scale person re-identiﬁcation.In ECCV, 2016.  L. Zheng, L. Shen, L. Tian, S. Wang, J. Wang, and Q. Tian. Scalableperson re-identiﬁcation: A benchmark. In ICCV, 2015.  L. Zheng, Y. Yang, and A. G. Hauptmann. Person re-identiﬁcation:Past, present and future. arXiv, 2016.  W. Zheng, S. Gong, and T. Xiang. Person re-identiﬁcation by proba-bilistic relative distance comparison. In CVPR, 2011.  Z. Zheng, L. Zheng, and Y. Yang. Unlabeled samples generated bygan improve the person re-identiﬁcation baseline in vitro. In ICCV,2017.  Z. Zhong, L. Zheng, D. Cao, and S. Li. Re-ranking person re-identiﬁcation with k-reciprocal encoding. In CVPR, 2017.  X. Zhu, X. Jing, F. Wu, and H. Feng. Video-based person re-identiﬁcation by simultaneously learning intra-video and inter-videodistance metrics. In IJCAI, 2016.