Multi-level max-margin analysis for semantic classification of satellite images

详细信息    查看全文

• 作者：Fan Hu (1) (2)
  Gui-Song Xia (2)
  Hong Sun (1)
  
  1. Electronic Information School ; Wuhan University ; Wuhan ; 430072 ; Hubei ; China
  2. State Key Laboratory of Information Engineering in Surveying ; Mapping and Remote Sensing ; Wuhan University ; Wuhan ; 430079 ; Hubei ; China
  
• 关键词：satellite image classification ; topic model ; maximum entropy discrimination latent Dirichlet allocation ; large margin nearest neighbor classifier ; multi ; level max ; margin ; TP 7
• 刊名：Wuhan University Journal of Natural Sciences
• 出版年：2015
• 出版时间：February 2015
• 年：2015
• 卷：20
• 期：1
• 页码：47-54
• 全文大小：982 KB
• 参考文献：1. Xia G S, Yang W, Delon J, / et al. Structural indexing of high-resolution satellite images [C]// / ISPRS TC VII / Symposium-100 / Years ISPRS. Vienna: ISPRS Press, 2010:298鈥?04.
  2. Hofmann T. Probabilistic latent semantic analysis [C]// / Proceedings of the 22 / nd Annual International ACM SIGIR Conference on Research and Development in Information Retrieval. Berkeley: ACM Press, 1999: 50鈥?7.
  3. Blei D M, Ng A Y, Jordan M I. Latent dirichlet allocation [J]. / The Journal of Machine Learning Research, 2003, 3: 993鈥?022.
  4. Hu F, Yang W, Chen J, / et al. Tile-level annotation of satellite images using multi-level max-margin discriminative random field [J]. / Remote Sensing, 2013, 5,(5): 2275鈥?291. CrossRef
  5. Blei D M, McAuliffe J D. Supervised topic models [C] // / Advances in Neural Information Processing Systems. Vancouver: NIPS Foundation Press, 2008: 121鈥?28.
  6. Lacoste-Julien S, Sha F, Jordan M I. DiscLDA: Discriminative learning for dimensionality reduction and classification[ C]// / Advances in Neural Information Processing Systems. Vancouver: NIPS Foundation Press, 2008: 897鈥?04.
  7. Zhu J, Ahmed A, Xing E P. MedLDA: Maximum margin supervised topic models for regression and classification[ C]// / Proceedings of the 26 / th Annual International Conference on Machine Learning. Montreal: ACM Press, 2009: 1257鈥?264.
  8. Wang Y, Mori G. Max-margin latent dirichlet allocation for image classification and annotation [C]// / British Machine Vision Conference (BMVC). Dundee: BMVA Press, 2011: 1鈥?1.
  9. Verbeek J, Triggs B. Region classification with markov field aspect models [C]_// / Computer Vision and Pattern Recognition (CVPR鈥?007). Minnesota: IEEE Press, 2007: 1鈥?.
  10. Weinberger K Q, Saul L K. Distance metric learning for large margin nearest neighbor classification [J]. / The Journal of Machine Learning Research, 2009, 10: 207鈥?44.
  11. Yang Y, Newsam S. Bag-of-visual-words and spatial extensions for land-use classification [C]// / Proceedings of the 18 / th SIGSPATIAL International Conference on Advances in Geographic Information Systems. San Jose: ACM Press, 2010: 270鈥?79.
  
• 刊物类别：Mathematics and Statistics
• 刊物主题：Mathematics
  Computer Science, general
  Physics
  Life Sciences
  Chinese Library of Science
  
• 出版者：Wuhan University, co-published with Springer
• ISSN：1993-4998

文摘

The performance of scene classification of satellite images strongly relies on the discriminative power of the low-level and mid-level feature representation. This paper presents a novel approach, named multi-level max-margin analysis (M3DA) for semantic classification for high-resolution satellite images. In our M3DA model, the maximum entropy discrimination latent Dirichlet allocation (MedLDA) model is applied to learn the topic-level features first, and then based on a bag-of-words representation of low-level local image features, the large margin nearest neighbor (LMNN) classifier is used to optimize a multiple soft label composed of word-level features (generated by SVM classifier) and topic-level features. The categorization performances on 21-class land-use dataset have demonstrated that the proposed model in multi-level max-margin scheme can distinguish different categories of land-use scenes reasonably.