Scene categorization based on local–global feature fusion and multi-scale multi-spatial resolution encoding

详细信息    查看全文

• 作者：Jianzhao Qin ; Fuqin Deng ; Nelson H. C. Yung
• 关键词：Scene categorization ; Local–global feature fusion ; Multi ; scale multi ; spatial resolution encoding
• 刊名：Signal, Image and Video Processing
• 出版年：2014
• 出版时间：December 2014
• 年：2014
• 卷：8
• 期：1-supp
• 页码：145-154
• 全文大小：1,046 KB
• 参考文献：1. Blei, DM, Ng, AY, Jordan, MI (2003) Latent dirichlet allocation. J. Mach. Learn. Res. 3: pp. 944937
  2. Bosch, A., Zisserman, A., Munoz, X.: Scene classification via plsa. In: ECCV 2006, pp. 517-30 (2006)
  3. Bosch, A, Zisserman, A, Muoz, X (2008) Scene classification using a hybrid generative/discriminative approach. IEEE Trans. Pattern Anal. Mach. Intell. 30: pp. 712-727 class="external" href="http://dx.doi.org/10.1109/TPAMI.2007.70716" target="_blank" title="It opens in new window">CrossRef
  4. Boureau, Y.L., Bach, F., LeCun, Y., Ponce, J.: Learning mid-level features for recognition. In: IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2010, pp. 2559-566 (2010)
  5. Fei-Fei, L, Perona, P (2005) A bayesian hierarchical model for learning natural scene categories. IEEE Comput. Soc. Conf. Comput. Vis. Pattern Recognit. 2: pp. 524-531
  6. Gehler, P., Nowozin, S.: On feature combination for multiclass object classification. In: IEEE 12th International Conference on Computer Vision, 2009 , pp. 221-28 (2009)
  7. Hofmann, T (2001) Unsupervised learning by probabilistic latent semantic analysis. Mach. Learn. 42: pp. 177-196 class="external" href="http://dx.doi.org/10.1023/A:1007617005950" target="_blank" title="It opens in new window">CrossRef
  8. Kwitt, R., Vasconcelos, N., Rasiwasia, N.: Scene recognition on the semantic manifold. In: Proceedings of the 12th European Conference on Computer Vision—Volume Part IV. ECCV-2, pp. 359-72. Springer, Berlin (2012)
  9. Lazebnik, S., Schmid, C., Ponce, J.: Beyond bags of features: spatial pyramid matching for recognizing natural scene categories. In: IEEE Computer Society Conference on Computer Vision and Pattern Recognition, 2006 , vol. 2, pp. 2169-178 (2006)
  10. Lee, J.J.: Libpmk: a pyramid match toolkit. Tech. Rep. MIT-CSAIL-TR-2008-17, MIT Computer Science and Artificial Intelligence Laboratory (2008)
  11. Li, T, Mei, T, Kweon, IS, Hua, XS (2011) Contextual bag-of-words for visual categorization. IEEE Trans. Circuits Syst. Video Technol. 21: pp. 381-392 class="external" href="http://dx.doi.org/10.1109/TCSVT.2010.2041828" target="_blank" title="It opens in new window">CrossRef
  12. Lian, X.C., Li, Z., Lu, B.L., Zhang, L.: Max-margin dictionary learning for multiclass image categorization. In: Daniilidis, K., Maragos, P., Paragios, N. (eds.) Computer Vision ECCV 2010. Lecture Notes in Computer Science, vol. 6314, pp. 157-70. Springer, Berlin (2010)
  13. Lowe, D.G.: Object recognition from local scale-invariant features. In: The Proceedings of the Seventh IEEE International Conference on Computer Vision, vol. 2, pp. 1150-157 (1999)
  14. Mahbub, U, Imtiaz, H, Ahad, MAR (2014) Action recognition based on statistical analysis from clustered flow vectors. Signal Image Video Process. 8: pp. 243-253 class="external" href="http://dx.doi.org/10.1007/s11760-013-0533-3" target="_blank" title="It opens in new window">CrossRef
  15. Ojala, T, Pietik?inen, M, M?enp??, T (2002) Multiresolution gray-scale and rotation invariant texture classification with local binary patterns. IEEE Trans. Pattern Anal. Mach. Intell. 24: pp. 971-987 class="external" href="http://dx.doi.org/10.1109/TPAMI.2002.1017623" target="_blank" title="It opens in new window">CrossRef
  16. Oliva, A, Torralba, A (2001) Modeling the shape of the scene: a holistic representation of the spatial envelope. Int. J. Comput. Vis. 42: pp. 145-175 class="external" href="http://dx.doi.org/10.1023/A:1011139631724" target="_blank" title="It opens in new window">CrossRef
  17. Pandey, M., Lazebnik, S.: Scene recognition and weakly supervised object localization with deformable part-based models. In: IEEE International Conference on Computer Vision (ICCV), 2011 pp. 1307-314 (2011). doi:class="a-plus-plus non-url-ref">10.1109/ICCV.2011.6126383
  18. Qin, J, Yung, NHC (2009) Scene categorization with multi-scale category-specific visual words. Opt. Eng. 48: pp. 047 class="external" href="http://dx.doi.org/10.1117/1.3115471" target="_blank" title="It opens in new window">CrossRef
  19. Qin, J, Yung, NHC (2010) Scene categorization via contextual visual words. Pattern Recognit. 43: pp. 1874-1888 class="external" href="http://dx.doi.org/10.1016/j.patcog.2009.11.009" target="_blank" title="It opens in new window">CrossRef
  20. Qin, J, Yung, NHC (2012) Feature fusion within local region using localized maximum-margin learning for scene categorization. Pattern Recognit. 45: pp. 1671-1683 class="external" href="http://dx.doi.org/10.1016/j.patcog.2011.09.027" target="_blank" title="It opens in new window">CrossRef
  21. Quattoni, A., Torralba, A.: Recognizing indoor scenes. In: IEEE Conference on Computer Vision and Pattern Recognition, 2009. CVPR 2009, pp. 413-20 (2009). doi:class="a-plus-plus non-url-ref">10.1109/CVPR.2009.5206537
  22. Siagian, C., Itti, L
• 刊物类别：Engineering
• 刊物主题：Signal,Image and Speech Processing
  Image Processing and Computer Vision
  Computer Imaging, Vision, Pattern Recognition and Graphics
  Multimedia Information Systems
  
• 出版者：Springer London
• ISSN：1863-1711

文摘

With the bag-of-contextual-visual-word (BOCVW) models, we propose a scene categorization method based on local–global feature fusion and multi-scale multi-spatial resolution encoding. First, the performances of the BOCVW models belonging to different features are mutually reinforced by fusing other types of features within local regions. Then, the spatial configuration information is explored using a multi-scale multi-spatial resolution encoding approach. Furthermore, these encoded BOCVW models are globally fused using an improved maximum-margin optimization strategy, which considers the margin between input vectors of different categories and the diameter of the smallest ball containing feature vectors simultaneously. The proposed method has been evaluated on three scene categorization datasets consisting of scene categories 8, 15, and 67, respectively. And its effectiveness has been verified by these experimental results.