access icon free Enhancing person re-identification by late fusion of low-, mid- and high-level features

© The Institution of Engineering and Technology
Received 16/12/2016, Accepted 02/07/2017, Revised 17/05/2017, Published 07/07/2017

Person re-identification is the process of finding people across different cameras. In this process, focus often lies in developing strong feature descriptors or a robust metric learning algorithm. While the two aspects are the most important steps in order to secure a high performance, a less explored aspect is late fusion of complementary features. For this purpose, this study proposes a late fusing scheme that, based on an experimental analysis, combines three systems that focus on extracting features and provide supervised learning on different abstraction levels. To analyse the behaviour of the proposed system, both rank aggregation and score-level fusion are applied. The authors’ proposed fusion scheme increases results on both small and large datasets. Experimental results on VIPeR show accuracies 5.43% higher than related systems, while results on PRID450S and CUHK01 increase state-of-the-art results by 10.94 and 14.84%, respectively. Furthermore, a cross-dataset test shows an increased rank-1 accuracy of 28.26% when training on CUHK02 and testing on VIPeR. Finally, an analysis of the late fusion shows aggregation to be better when individual results are unequally distributed within top-10 while score-level fusion provides better results when two individual results lie within top-5 while the last lies outside top-10.

Inspec keywords: learning (artificial intelligence); image fusion; feature extraction

Other keywords: mid-level features; high-level features; feature extraction; low-level features; strong feature descriptors; person re-identification; robust metric learning algorithm; VIPeR; score-level fusion; supervised learning; late fusion; CUHK01; late fusing scheme; PRID450S; abstraction levels; rank aggregation

Subjects: Sensor fusion; Image recognition; Knowledge engineering techniques; Computer vision and image processing techniques

References

    1. 1)
      • 14. Pantic, M., Rothkrantz, L.J.: ‘Toward an affect-sensitive multimodal human-computer interaction’, Proc. IEEE, 2003, 91, (9), pp. 1370–1390.
    2. 2)
      • 27. Wu, S., Chen, Y.-C., Zheng, W.-S.: ‘An enhanced deep feature representation for person reidentification’. Proc. WACV, 2016, pp. 1–8.
    3. 3)
      • 38. Stuart, J.M., Segal, E., Koller, D., et al: ‘A gene-coexpression network for global discovery of conserved genetic modules’, Science, 2003, 302, (5643), pp. 249–255.
    4. 4)
      • 16. Dwork, C., Kumar, R., Naor, M., et al: ‘Rank aggregation methods for the web’. Proc. WWW, 2001, pp. 613–622.
    5. 5)
      • 7. Bazzani, L., Cristani, M., Murino, V.: ‘Symmetry-driven accumulation of local features for human characterization and re-identification’, Comput. Vis. Image Underst., 2013, 117, (2), pp. 130–144.
    6. 6)
      • 9. Zhao, R., Ouyang, W., Wang, X.: ‘Learning mid-level filters for person re-identification’. Proc. CVPR, 2014, pp. 144–151.
    7. 7)
      • 29. Martinel, N., Micheloni, C., Foresti, G.L.: ‘A pool of multiple person re-identification experts’, Pattern Recognit. Lett., 2016, 71, pp. 23–30.
    8. 8)
      • 15. Kittler, J., Hatef, M., Duin, R.P., et al: ‘On combining classifiers’, IEEE Trans. Pattern Anal. Mach. Intell., 1998, 20, (3), pp. 226–239.
    9. 9)
      • 20. Prosser, B., Zheng, W.-S., Gong, S., et al: ‘Person re-identification by support vector ranking’. Proc. BMVC, 2010, pp. 21.1–21.11.
    10. 10)
      • 19. Yang, Y., Yang, J., Yan, J., et al: ‘Salient color names for person reidentification’. Proc. ECCV, 2014, pp. 536–551.
    11. 11)
      • 18. Liao, S., Hu, Y., Zhu, X., et al: ‘Person re-identification by local maximal occurrence representation and metric learning’. Proc. CVPR, 2015, pp. 2197–2206.
    12. 12)
      • 11. Krizhevsky, A., Sutskever, I., Hinton, G.E.: ‘Imagenet classification with deep convolutional neural networks’. Proc. NIPS, 2012, pp. 1097–1105.
    13. 13)
      • 13. Xiong, F., Gou, M., Camps, O., et al: ‘Person re-identification using kernel-based metric learning methods’. Proc. ECCV, 2014, pp. 1–16.
    14. 14)
      • 12. Koestinger, M., Hirzer, M., Wohlhart, P., et al: ‘Large scale metric learning from equivalence constraints’. Proc. CVPR, 2012, pp. 2288–2295.
    15. 15)
      • 35. Li, S., Shao, M., Fu, Y.: ‘Cross-view projective dictionary learning for person re-identification’. Proc. AAAI, 2015, pp. 2155–2161.
    16. 16)
      • 17. Zhao, R., Ouyang, W., Wang, X.: ‘Unsupervised salience learning for person re-identification’. Proc. CVPR, 2013, pp. 3586–3593.
    17. 17)
      • 8. Bak, S., Brémond, F.: ‘Re-identification by covariance descriptors’, in Gong, S., Cristani, M., Yan, S., et al (Eds.): ‘Person re-identification, volume 1 of advances in computer vision and pattern recognition’ (Springer, London, 2014, 1st edn.), Ch. 4, pp. 71–91.
    18. 18)
      • 33. Jobson, D.J., Rahman, Z.-u., Woodell, G.A.: ‘A multiscale retinex for bridging the gap between color images and the human observation of scenes’, IEEE Trans. Image Process., 1997, 6, (7), pp. 965–976.
    19. 19)
      • 30. Liu, X., Wang, H., Wu, Y., et al: ‘An ensemble color model for human reidentification’. Proc. WACV, 2015, pp. 868–875.
    20. 20)
      • 34. Liao, S., Zhao, G., Kellokumpu, V., et al: ‘Modeling pixel process with scale invariant local patterns for background subtraction in complex scenes’. Proc. CVPR, 2010, pp. 1301–1306.
    21. 21)
      • 39. Kolde, R., Laur, S., Adler, P., et al: ‘Robust rank aggregation for gene list integration and meta-analysis’, Bioinformatics, 2012, 28, (4), pp. 573–580.
    22. 22)
      • 4. Li, W., Zhao, R., Xiao, T., et al: ‘Deepreid: deep filter pairing neural network for person re-identification’. Proc. CVPR, 2014, pp. 152–159.
    23. 23)
      • 40. Roth, P.M., Hirzer, M., Köstinger, M.: ‘Mahalanobis distance learning for person re-identification’, Gong, S., Cristani, M., Yan, S., et al (Eds.): ‘Person re-identification, volume 1 of advances in computer vision and pattern recognition’ (Springer, London, 2014, 1st edn.), ch. 12, pp. 247–267.
    24. 24)
      • 32. de Carvalho Prates, R.F., Schwartz, W.R.: ‘Cbra: color-based ranking aggregation for person re-identification’. Proc. ICIP, 2015, pp. 1975–1979.
    25. 25)
      • 24. Ahmed, E., Jones, M., Marks, T.K.: ‘An improved deep learning architecture for person reidentification’. Proc. CVPR, 2015, pp. 3908–3916.
    26. 26)
      • 36. Chen, Y.-C., Zheng, W.-S., Lai, J.: ‘Mirror representation for modeling view-specific transform in person re-identification’. Proc. IJCAI, 2015, pp. 3402–3408.
    27. 27)
      • 37. de Carvalho Prates, R.F., Schwartz, W.R.: ‘Appearance-based person re-identification by intra-camera discriminative models and rank aggregation’. Proc. ICB, 2015, pp. 65–72.
    28. 28)
      • 5. Zheng, L., Shen, L., Tian, L., et al: ‘Scalable person re-identification: a benchmark’. Proc. ICCV, 2015, pp. 1116–1124.
    29. 29)
      • 26. Paisitkriangkrai, S., Shen, C., van den Hengel, A.: ‘Learning to rank in person re-identification with metric ensembles’. Proc. CVPR, 2015, pp. 1846–1855.
    30. 30)
      • 21. Zheng, W.-S., Gong, S., Xiang, T.: ‘Person re-identification by probabilistic relative distance comparison’. Proc. CVPR, 2011, pp. 649–656.
    31. 31)
      • 42. Chen, S.-Z., Guo, C.-C., Lai, J.-H.: ‘Deep ranking for person re-identification via joint representation learning’, IEEE Trans. Image Process., 2016, 25, (5), pp. 2353–2367.
    32. 32)
      • 25. Yi, D., Lei, Z., Liao, S., et al: ‘Deep metric learning for person re-identification’. Proc. ICPR, 2014, pp. 34–39.
    33. 33)
      • 23. Liu, X., Song, M., Tao, D., et al: ‘Semi-supervised coupled dictionary learning for person re-identification’. Proc. CVPR, 2014, pp. 3550–3557.
    34. 34)
      • 28. Zheng, L., Wang, S., Tian, L., et al: ‘Query-adaptive late fusion for image search and person re-identification’. Proc. CVPR, 2015, pp. 1741–1750.
    35. 35)
      • 31. Ye, M., Chen, J., Leng, Q., et al: ‘Coupled-view based ranking optimization for person re-identification’. Proc. MMM, 2015, pp. 105–117.
    36. 36)
      • 6. Gray, D., Tao, H.: ‘Viewpoint invariant pedestrian recognition with an ensemble of localized features’. Proc. ECCV, 2008, pp. 262–275.
    37. 37)
      • 3. Schumann, A., Monari, E.: ‘A soft-biometrics dataset for person tracking and reidentification’. Proc. AVSS, 2014, pp. 193–198.
    38. 38)
      • 22. Zheng, L., Shen, L., Tian, L., et al: ‘Person re-identification meets image search’. CoRR, 2015, pp. 4321–4330, abs/1502.02171. Available at http://arxiv.org/abs/1502.02171, retrieved from: http://arxiv.org/abs/1502.02171.
    39. 39)
      • 1. Liu, Z., Zhang, Z., Wu, Q., et al: ‘Enhancing person re-identification by integrating gait biometric’, Neurocomputing, 2015, 168, pp. 1144–1156.
    40. 40)
      • 2. DeCann, B., Ross, A.: ‘Modelling errors in a biometric re-identification system’, IET Biometrics, 2015, 4, (4), pp. 209–219.
    41. 41)
      • 10. Lazebnik, S., Schmid, C., Ponce, J.: ‘Beyond bags of features: spatial pyramid matching for recognizing natural scene categories’. Proc. CVPR, 2006, vol. 2, pp. 2169–2178.
    42. 42)
      • 41. Li, W., Zhao, R., Wang, X.: ‘Human reidentification with transferred metric learning’. Proc. ACCV, 2012, pp. 31–44.
http://iet.metastore.ingenta.com/content/journals/10.1049/iet-bmt.2016.0200
Loading

Related content

content/journals/10.1049/iet-bmt.2016.0200
pub_keyword,iet_inspecKeyword,pub_concept
6
6
Loading