Pedestrian re-identification technology has become the current research focus due to its wide range of applications. This study conducted cross dataset pedestrian re-identification to solve the problem that the single dataset’s difficulty for simulating the actual situation and its poor generalization ability. Deep learning has made remarkable achievements in the fields of machine learning recently, so the deep learning technology is integrated into cross datasets pedestrian re-identification system. Here we improved the three-layer convolutional neural network (CNN) structure proposed by Yang Hu in Asia Conference on Computer Vision (ACCV), 2014. The Batch Normalization (BN) layer has been added to reduce the over-fitting degree during training period and the adjusted cosine similarity algorithm is used for pedestrian feature match to solve the defect of cosine similarity algorithm. Finally we implemented the entire cross dataset pedestrian re-identification system and got the experimental results. The Shinpuhkan2014dataset was chosen as training set. We compared the training results before and after adding BN layer and found that test accuracy increased, test loss decreased and over-fitting phenomenon eased. The VIPeR and i_LIDS datasets were chosen as test sets. We evaluated the effects on VIPeR and i_LIDS based on the CNN model that training on Shinpuhkan2014dataset. The cumulative matching rate rank5 increased by 1.7% on VIPeR dataset compared with the current level, the rank10 and rank20 also increased. And the cumulative matching rate rank1 increased by 1.8% on i_LIDS dataset compared with the current level, the rank5 and rank10 also increased.

Cross dataset, Convolutional neural network, Batch normalization, Adjusted cosine similarity

[1] Tulsi Jain, Kushagra Agarwal, Ronil Pancholia, “Sentiment Analysis Based on a Deep Stochastic Network and Active Learning”, International Journal of Computer Sciences and Engineering (IJCSE), Vol.5, Issue.9, pp.1-6, 2017.
[2] Taigman, Y., Yang, M., Ranzato, M., Wolf, L, “DeepFace: Closing the Gap to Human-Level Performance in Face Verification”, IEEE Conference on Computer Vision and Pattern Recognition, pp.1701-1708, 2014.
[3] Krizhevsky A, Sutskever I, Hinton G E, “ImageNet classification with deep convolutional neural networks”, International Conference on Neural Information Processing Systems. Curran Associates Inc. pp.1097-1105, 2012.
[4] Kawanishi, Y., Yang, W., Mukunoki, M., Minoh, M., “Shinpuhkan2014: A Multi-Camera Pedestrian Dataset for Tracking People across Multiple Cameras”, The Korea-Japan Joint Workshop on Frontiers of Computer Vision, FCV, 2014.
[5] Gray D, Brennan S, Tao H, “Evaluating appearance models for recognition, reacquisition, and tracking”, Vol. 3, Issue. 5, pp.1-7, 2007.
[6] Zheng W S, Gong S, Xiang T, “Person re-identification by probabilistic relative distance comparison”, In Computer Vision and Pattern Recognition(CVPR), IEEE,Vol. 42, pp.649-656,2011.
[7] Weinberger, K., Blitzer, J., Saul, L., “Distance metric learning for large margin nearest neighbor classification”, Advances in neural information processing systems 18, pp.1473-1480, 2006.
[8] Dikmen, M., Akbas, E., Huang, T. S., Ahuja, N., “Pedestrian recognition with a learned metric”, Lecture Notes in Computer Science, 6495 pp.501-512, 2010.
[9] Davis J V, Kulis B, Jain P, Sra S, Dhillon IS, “Information-theoretic metric learning”, International Conference on Machine Learning, ACM, Vol. 227, pp.209-216, 2007.
[10] Zheng W S, Gong S, Xiang T, “Person re-identification by probabilistic relative distance comparison”, In Computer Vision and Pattern Recognition(CVPR), Vol. 42, pp.649-656, 2011.
[11] Köstinger M, Hirzer M, Wohlhart P, Roth PM, Bischof H, “Large scale metric learning from equivalence constraints”. IEEE Conference on Computer Vision and Pattern Recognition. IEEE Computer Society, pp.2288-2295, 2012.
[12] Li, Z., Chang, S., Liang, F., Huang, T.S., Cao, L., Smith, J.R, “Learning locally adaptive decision functions for person verification”, In Computer Vision and Pattern Recognition (CVPR), IEEE Conference, pp.3610-3617, 2013.
[13] Hu, Yang, Liao, Shengcai, Lei, Zhen, Yi, Dong, Li, Stan, “Exploring Structural Information and Fusing Multiple Features for Person Re-identification”, Computer Vision and Pattern Recognition Workshops, IEEE, pp.794-799, 2013.
[14] Gheissari N, Sebastian T B, Hartley R, “Person Re-identification Using Spatiotemporal Appearance”, Computer Vision and Pattern Recognition, Computer Society Conference on IEEE, pp.1528-1535, 2006.
[15] Hamdoun O, Moutarde F, Stanciulescu B, Steux B, “Person re-identification in multi-camera system by signature based on interest point descriptors collected on short video sequences”, International Conference on Distributed Smart Cameras, IEEE, pp.1-6, 2008.
[16] Wang X, Doretto G, Sebastian T, Rittscher J, Tu P, “Shape and Appearance Context Modeling”, International Conference on Computer Vision, IEEE, pp.1-8, 2007.
[17] Farenzena, M., Bazzani, L., Perina, A., Murino, V., & Cristani, M., “Person re-identification by symmetry-driven accumulation of local features”, IEEE Computer Vision and Pattern Recognition, Vol.23, pp.2360-2367,2010.
[18] Ma B, Su Y, “Local descriptors encoded by fisher vectors for person re-identification”, International Conference on Computer Vision. Springer-Verlag, pp.413-422, 2012.
[19] Dong S C, Cristani M, Stoppa M, Bazzani L, Murino V, “Custom Pictorial Structures for Re-identification”, British Machine Vision Conference(BMVC), Vol. 68, pp.1-11, 2011.
[20] Zhao R, Ouyang W, Wang X, “Unsupervised Salience Learning for Person Re-identification”, Computer Vision and Pattern Recognition, IEEE, pp.3586-3593, 2013.
[21] Zhao R, Ouyang W, Wang X, “Person Re-identification by Salience Matching”, International Conference on Computer Vision, IEEE, pp.2528-2535, 2014.
[22] Liu Y, Shao Y, Sun F, “Person re-identification based on visual saliency”, Intelligent Systems Design and Applications (ISDA) , Vol. 13,Issue. 6, pp.884-889, 2012.
[23] Gray D, Tao H, “Viewpoint Invariant Pedestrian Recognition with an Ensemble of Localized Features”, Computer Vision - ECCV 2008, European Conference on Computer Vision, Marseille, France, Proceedings. DBLP, pp.262-275, October 12-18, 2008,.
[24] Prosser, B., Zheng, W.S., Gong, S., Xiang, T., Mary, Q., “Person re-identification by support vector ranking”, In British Machine Vision Conference (BMVC), Aberystwyth, UK:BMVA Press, Vol. 2, Issue. 5, pp.1-11, 2010.
[25] Li W, Wang X, “Locally Aligned Feature Transforms across Views”, Conference on Computer Vision and Pattern Recognition, IEEE Computer Society, pp.3594-3601, 2013.
[26] Liu C, Chen C L, Gong S, Wang G, “POP: Person Re-identification Post-rank Optimization”, IEEE International Conference on Computer Vision, IEEE Computer Society, pp.441-448, 2013.
[27] Ma A J, Yuen P C, Li J, “Domain Transfer Support Vector Ranking for Person Re-identification without Target Camera Label Information”, IEEE International Conference on Computer Vision, IEEE, pp.3567-3574, 2014.
[28] Yi D, Lei Z, Liao S, Li SZ, “Deep Metric Learning for Person Re-identification”. International Conference on Pattern Recognition, IEEE, pp.34-39, 2014.
[29] Yang Hu, Dong Yi, Shengcai Liao, Zhen Lei, Stan Z, “Cross Dataset Person Re-identification”, Institute of Automation, Chinese Academy of Sciences (CASIA), ACCV, 9010, pp.650-664, 2014.
[30] Ioffe S, Szegedy C, “Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift”, arXiv preprint arXiv:1502.03167, pp.448-456, 2015.
[31] Globerson A, Roweis S T, “Metric Learning by Collapsing Classes”, Advances in Neural Information Processing Systems, Vol. 18, pp.451-458, 2006.
[32] Gray D, Tao H, “Viewpoint Invariant Pedestrian Recognition with an Ensemble of Localized Features”, Computer Vision - ECCV 2008, European Conference on Computer Vision, Marseille, France , Proceedings. DBLP, pp.262-275, October 12-18, 2008.
[33] Xing E P, Ng A Y, Jordan M I, Russell S, “Distance metric learning, with application to clustering with side-information”, International Conference on Neural Information Processing Systems, MIT Press, pp.521-528, 2002.