Advertisement

Multi-center convolutional descriptor aggregation for image retrieval

  • Jie Zhu
  • Shufang Wu
  • Hong Zhu
  • Yan Li
  • Li Zhao
Original Article
  • 28 Downloads

Abstract

Recent works have demonstrated that the convolutional descriptor aggregation can provide state-of-the-art performance for image retrieval. In this paper, we propose a multi-center convolutional descriptor aggregation (MCDA) method to produce global image representation for image retrieval. We first present a feature map center selection method to eliminate the background information in the feature maps. We then propose the channel weighting and spatial weighting schemes based on the centers to boost the effect of the features on the object. Finally, the weighted convolutional descriptors are aggregated to represent images. Experiments demonstrate that MCDA can produce state-of-the-art retrieval performance, and the generated activation map is also effective for object localization.

Keywords

Multi-center Descriptor aggregation Feature map Feature weighting 

Notes

Acknowledgements

This work is supported by the National Natural Science Foundation of China (Grant 61772344 and Grant 61732011), in part by the Natural Science Foundation of SZU (Grant 827-000140, Grant 827-000230, and Grant 2017060), the National Social Science Foundation of China (17BTQ068), the Youth Foundation of Education Bureau of Hebei Province (Grant QN2015099), China Postdoctoral Science Foundation funded project (Grant 2017M621078). The funding project of midwest colleges and universities promoting comprehensive strength of Hebei University.

References

  1. 1.
    Krizhevsky A, Sutskever I, Hinton GE(2012) Imagenet classification with deep convolutional neural networks. In: Annual conference on neural information processing systems, pp 1097–1105Google Scholar
  2. 2.
    Cui L, Yang S, Chen F et al (2018) A survey on application of machine learning for internet of things. Int J Mach Learn Cybernet 9(8):1399–1417CrossRefGoogle Scholar
  3. 3.
    Banharnsakun A (2018) Towards improving the convolutional neural networks for deep learning using the distributed artificial bee colony method. Int J Mach Learn Cybern.  https://doi.org/10.1007/s13042-018-0811-z CrossRefGoogle Scholar
  4. 4.
    Cui Y, Jiang J, Lai Z et al (2018) Supervised discrete discriminant hashing for image retrieval. Pattern Recogn 78:79–90CrossRefGoogle Scholar
  5. 5.
    Kalantidis Y, Mellina C, Osindero S (2016) Cross-dimensional weighting for aggregated deep convolutional features. European conference on computer vision. Springer, Cham, pp 685–701Google Scholar
  6. 6.
    Wei XS, Luo JH, Wu J (2017) Selective convolutional descriptor aggregation for fine-grained image retrieval. IEEE Trans Image Process 26(6):2868–2881MathSciNetCrossRefGoogle Scholar
  7. 7.
    Ng JYH, Yang F, Davis LS (2015) Exploiting local features from deep networks for image retrieval. arXiv preprint: 1504.05133Google Scholar
  8. 8.
    Jegou H, Douze M, Schmid C (2008) Hamming embedding and weak geometric consistency for large scale image search. European conference on computer vision, pp 304–317Google Scholar
  9. 9.
    Philbin J, Chum O, Isard M (2008) Lost in quantization: Improving particular object retrieval in large scale image databases. International conference on computer vision and pattern recognition, pp 1–8Google Scholar
  10. 10.
    Philbin J, Chum O, Isard M (2007) Object retrieval with large vocabularies and fast spatial matching. International conference on computer vision and pattern recognition, pp 1–8Google Scholar
  11. 11.
    Wang XZ, Wang R, Xu C (2018) Discovering the relationship between generalization and uncertainty by incorporating complexity of classification. IEEE Trans Cybern 48(2):703–715MathSciNetCrossRefGoogle Scholar
  12. 12.
    Wang R, Kwong S, Wang XZ et al (2015) Segment based decision tree induction with continuous valued attributes. IEEE Trans Cybern 45(7):1262–1275CrossRefGoogle Scholar
  13. 13.
    Wang R, Chow CY, Kwong S (2016) Ambiguity-based multiclass active learning. IEEE Trans Fuzzy Syst 24(1):242–248CrossRefGoogle Scholar
  14. 14.
    Wang R, Wang XZ, Kwong S et al (2017) Incorporating diversity and informativeness in multiple-instance active learning. IEEE Trans Fuzzy Syst 25(6):1460–1475CrossRefGoogle Scholar
  15. 15.
    Jégou H, Douze M, Schmid C (2010) Aggregating local descriptors into a compact image representation. International conference on computer vision and pattern recognition, pp 3304–3311Google Scholar
  16. 16.
    Sánchez J, Perronnin F, Mensink T et al (2013) Image classification with the fisher vector: theory and practice. Int J Comput Vis 105(3):222–245MathSciNetCrossRefGoogle Scholar
  17. 17.
    Babenko A, Slesarev A, Chigorin A (2014) Neural codes for image retrieval. European conference on computer vision, pp 584–599Google Scholar
  18. 18.
    Razavian AS, Azizpour H, Sullivan J, Carlsson S (2014) CNN features off-the-shelf: an astounding baseline for recognition. International conference on computer vision and pattern recognition workshops, pp 512–519Google Scholar
  19. 19.
    Gong Y, Wang L, Guo R (2014) Multi-scale orderless pooling of deep convolutional activation features. European conference on computer vision, pp 392–407Google Scholar
  20. 20.
    Babenko A, Lempitsky V (2015) Aggregating local deep features for image retrieval. IEEE international conference on computer vision, pp 1269–1277Google Scholar
  21. 21.
    Tolias G, Sicre R, Jégou H (2015) Particular object retrieval with integral max-pooling of CNN activations. arXiv preprint 1511.05879Google Scholar
  22. 22.
    Radenović F, Tolias G, Chum O (2016) CNN image retrieval learns from BoW: unsupervised fine-tuning with hard examples. European conference on computer vision, pp 3–20Google Scholar
  23. 23.
    Liu Z, Li J, Shen Z (2017) Learning efficient convolutional networks through network slimming. IEEE international conference on computer vision, pp 2755–2763Google Scholar
  24. 24.
    Boscaini D, Masci J, Rodolà E (2016) Learning shape correspondence with anisotropic convolutional neural networks. Adv Neural Inf Process Syst 3189–3197Google Scholar
  25. 25.
    Azizpour H, Sharif Razavian A, Sullivan J, Maki A, Carlsson S (2015) From generic to specific deep representations for visual recognition. International conference on computer vision and pattern recognition workshops, pp 36–45Google Scholar
  26. 26.
    Fu Z, Robles-Kelly A, Zhou J (2011) MILIS: Multiple instance learning with instance selection. IEEE Trans Pattern Anal Mach Intell 33(5):958–977CrossRefGoogle Scholar
  27. 27.
    Simonyan K, Zisserman A (2015) Very deep convolutional networks for large-scale image recognition. International conference on learning representations, pp 1–14Google Scholar
  28. 28.
    Razavian AS, Sullivan J, Carlsson S (2016) Visual instance retrieval with deep convolutional networks. ITE Trans Media Technol Appl 4(3):251–258CrossRefGoogle Scholar
  29. 29.
    Huang R, Zhang G, Chen J (2018) Semi-supervised discriminant Isomap with application to visualization, image retrieval and classification. Int J Mach Learn Cybern.  https://doi.org/10.1007/s13042-018-0809-6 CrossRefGoogle Scholar
  30. 30.
    Zhu Q, Yuan N, Guan D et al (2018) An alternative to face image representation and classification. Int J Mach Learn Cybern.  https://doi.org/10.1007/s13042-018-0802-0 CrossRefGoogle Scholar
  31. 31.
    Liu J, Liu W, Ma S et al (2018) Image-set based face recognition using K-SVD dictionary learning. Int J Mach Learn Cybern.  https://doi.org/10.1007/s13042-017-0782-5 CrossRefGoogle Scholar
  32. 32.
    Ding S, Zhang N, Zhang J et al (2017) Unsupervised extreme learning machine with representational features. Int J Mach Learn Cybernet 8(2):587–595CrossRefGoogle Scholar
  33. 33.
    Fang J, Xu X, Liu H et al (2018) Local receptive field based extreme learning machine with three channels for histopathological image classification. Int J Mach Learn Cybern.  https://doi.org/10.1007/s13042-018-0825-6 CrossRefGoogle Scholar

Copyright information

© Springer-Verlag GmbH Germany, part of Springer Nature 2018

Authors and Affiliations

  • Jie Zhu
    • 1
  • Shufang Wu
    • 2
  • Hong Zhu
    • 3
  • Yan Li
    • 4
  • Li Zhao
    • 3
  1. 1.Department of Information ManagementThe National Police University for Criminal JusticeBaodingChina
  2. 2.College of ManagementHebei UniversityBaodingChina
  3. 3.College of Computer Science and Software EngineeringShen Zhen UniversityShenzhenChina
  4. 4.School of Applied MathematicsBeijing Normal University ZhuhaiZhuhaiChina

Personalised recommendations