Asymmetric Semi-Supervised Boosting Scheme for Interactive Image Retrieval

ETRI Journal

关键词: boosting;    active learning;    semi-supervised learning;    support vector machines;    Interactive image retrieval;   

Others  :  1185899
DOI  :  10.4218/etrij.10.1510.0016

【 摘 要 】

Support vector machine (SVM) active learning plays a key role in the interactive content-based image retrieval (CBIR) community. However, the regular SVM active learning is challenged by what we call “the small example problem” and “the asymmetric distribution problem.” This paper attempts to integrate the merits of semi-supervised learning, ensemble learning, and active learning into the interactive CBIR. Concretely, unlabeled images are exploited to facilitate boosting by helping augment the diversity among base SVM classifiers, and then the learned ensemble model is used to identify the most informative images for active learning. In particular, a bias-weighting mechanism is developed to guide the ensemble model to pay more attention on positive images than negative images. Experiments on 5000 Corel images show that the proposed method yields better retrieval performance by an amount of 0.16 in mean average precision compared to regular SVM active learning, which is more effective than some existing improved variants of SVM active learning.

   

【 参考文献 】

• [1]X. Zhou and T.S. Huang, "Relevance Feedback in Image Retrieval: A Comprehensive Review," Multimedia Syst., vol. 8, no. 6, 2003, pp. 536-544.
• [2]Y. Liu et al., "A Survey of Content-Based Image Retrieval with High-Level Semantics," Pattern Recog., vol. 40, no. 1, 2007, pp. 262-282.
• [3]L. Zhang, F. Lin, and B. Zhang, "Support Vector Machine Learning for Image Retrieval," Proc. IEEE ICIP, 2001, pp. 721-724.
• [4]D.H. Kim et al., "Support Vector Machine Learning for Region-Based Image Retrieval with Relevance Feedback," ETRI J., vol. 29, no. 5, 2007, pp. 700-702.
• [5]S. Tong and E. Chang, "Support Vector Machine Active Learning for Image Retrieval," Proc. ACM Multimedia, 2001, pp. 107-118.
• [6]W. Jiang, G. Er, and Q. Dai, "Boost SVM Active Learning for Content-Based Image Retrieval," Proc. Asilomar Conf. Signals, Syst., Comput., 2003, pp. 1585-1589.
• [7]L. Wang, K.L. Chan, and Z. Zhang, "Bootstrapping SVM Active Learning by Incorporating Unlabelled Images for Image Retrieval," Proc. IEEE, CVPR, vol. 1, 2003, pp. 629-634.
• [8]S.C.H. Hoi et al., "Semi-supervised SVM Batch Mode Active Learning for Image Retrieval," Proc. IEEE CVPR, 2008, pp. 1-7.
• [9]D.C. Tao et al., "Asymmetric Bagging and Random Subspace for Support Vector Machines-Based Relevance Feedback in Image Retrieval," IEEE Trans. Pattern Anal. Mach. Intell., vol. 28, no. 7, 2006, pp. 1088-1099.
• [10]T.S. Huang et al., "Active Learning for Interactive Multimedia Retrieval," Proc. IEEE, vol. 96, no. 4, 2008, pp. 648-667.
• [11]V.N. Vapnik, Statistical Learning Theory, Wiley, 1998.
• [12]Z.H. Zhou, "Ensemble Learning," Encyclopedia of Biometrics, Springer, 2009, pp. 116-123.
• [13]O. Chapelle, B. Scholkipf, and A. Zien, Semi-supervised Learning, MIT Press, 2006.
• [14]Z.H. Zhou, "When Semi-supervised Learning Meets Ensemble Learning," Proc. Int. Workshop Multiple Classifier System, LNCS 5519, vol. 1, 2009, pp. 529-538.
• [15]P.K. Mallapragada et al., "SemiBoost: Boosting for Semi-Supervised Learning," IEEE Trans. Pattern Anal. Mach. Intell., vol. 31, no. 11, 2009, pp. 2000-2014.
• [16]L. Zheng et al., "Information Theoretic Regularization for Semi-Supervised Boosting," Proc. ACM SIGKDD, 2009, pp. 1017-1026.
• [17]Z.H. Zhou and M. Li, "Tri-Training: Exploiting Unlabeled Data Using Three Classifiers," IEEE Trans. Knowledge and Data Engineering, vol. 17, no. 11, 2005, pp. 1529-1541.
• [18]M. Li and Z.H. Zhou, "Classifier Ensemble with Unlabeled Data," CORR, vol. abs/0909.3593, 2009.
• [19]K. Bennett, A. Demiriz, and R. Maclin, "Exploiting Unlabeled Data in Ensemble Methods," Proc. ACM SIGKDD, 2002, pp. 289-296.
• [20]F. d'Alche-Buc, Y. Grandvalet, and C. Ambroise, "Semi-Supervised MarginBoost," NIPS, vol. 14, 2002, pp. 553-560.
• [21]N.V. Chawla, N. Japkowicz, and A. Kolcz, "Editorial: Special Issue on Learning from Imbalanced Data Sets," Proc. ACM SIGKDD Explorations, vol. 6, no. 1, 2004, pp. 1-6.
• [22]P. Viola, and M. Jones, "Fast and Robust Classification Using Asymmetric AdaBoost and a Detector Cascade," NIPS, vol. 14, 2002, pp. 1311-1318.
• [23]X.Y. Liu, J.X. Wu, and Z.H. Zhou, "Exploratory Undersampling for Class-Imbalance Learning," IEEE Trans. Syst., Man., Cybern. B, Cybern., vol. 39, no. 2, 2009, pp. 539-550.
• [24]C. Chang, and C. Lin, "LIBSVM: A Library for Support Vector Machines," 2001. Available: http://www.csie.ntu.edu.tw/~cjlin/ libsvm
• [25]D.P. Huijsmans and N. Sebe, "How to Complete Performance Graphs in Content-Based Image Retrieval: Add Generality and Normalize Scope," IEEE Trans. Pattern Anal. Mach. Intell., vol. 27, no. 2, 2005, pp. 245-251.