logo

SCIENTIA SINICA Informationis, Volume 49, Issue 2: 172-187(2019) https://doi.org/10.1360/N112018-00203

Semantic annotation of national cultural patterns based on dictionary learning

More info
  • ReceivedNov 29, 2018
  • AcceptedJan 15, 2019
  • PublishedFeb 21, 2019

Abstract

The national cultural pattern is a precious treasure of the Chinese nation. Semantic annotation and analysis of national cultural pattern is useful in cultural heritage and modern recreation. This article presents a multi-label dictionary learning algorithm called similar coefficient multi-label incoherent dictionary learning (SCMIDL) based on a multi-class dictionary learning algorithm. SCMIDL combines the incoherence of the dictionary and the similarity of coefficients, which significantly improves the performance of multi-label annotation. The superior annotation ability of the algorithm was confirmed on three kinds of national cultural pattern datasets constructed for evaluation purposes.


Funded by

北京市科委基金(D171100003717003)


References

[1] Gibaja E, Ventura S. Multi-label learning: a review of the state of the art and ongoing research. WIREs Data Min Knowl Discov, 2014, 4: 411-444 CrossRef Google Scholar

[2] Zhang M L, Zhou Z H. A Review on Multi-Label Learning Algorithms. IEEE Trans Knowl Data Eng, 2014, 26: 1819-1837 CrossRef Google Scholar

[3] Boutell M R, Luo J, Shen X. Learning multi-label scene classification. Pattern Recognition, 2004, 37: 1757-1771 CrossRef Google Scholar

[4] Read J, Pfahringer B, Holmes G. Classifier chains for multi-label classification. Mach Learn, 2011, 85: 333-359 CrossRef Google Scholar

[5] Fürnkranz J, Hüllermeier E, Loza Mencía E. Multilabel classification via calibrated label ranking. Mach Learn, 2008, 73: 133-153 CrossRef Google Scholar

[6] Tsoumakas G, Katakis I, Vlahavas I. Random k-Labelsets for Multilabel Classification. IEEE Trans Knowl Data Eng, 2011, 23: 1079-1089 CrossRef Google Scholar

[7] Zhang M L, Zhou Z H. ML-KNN: A lazy learning approach to multi-label learning. Pattern Recognition, 2007, 40: 2038-2048 CrossRef Google Scholar

[8] Clare A, King R D. Knowledge Discovery in Multi-label Phenotype Data. In: Proceedings of the 5th European Conference on Principles of Data Mining and Knowledge Discovery. Berlin: Springer, 2001. 2168: 42--53. Google Scholar

[9] Elisseeff A, Weston J. A kernel method for multi-labelled classification. In: Proceedings of International Conference on Neural Information Processing Systems: Natural and Synthetic. Cambridge: MIT Press, 2001. 681--687. Google Scholar

[10] Ghamrawi N, McCallum A. Collective multi-label classification. In: Proceedings of the 14th ACM international conference on Information and knowledge management. New York: ACM, 2005. 195--200. Google Scholar

[11] Berger M J. Large Scale Multi-Label Text Classification With Semantic Word Vectors. Technical report, Stanford University, 2015. Google Scholar

[12] Kurata G, Xiang B, Zhou B. Improved neural network-based multi-label classification with better initialization leveraging label co-occurrence. In: Proceedings of the 2016 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, 2016. 521--526. Google Scholar

[13] Chen G, Ye D, Xing Z, et al. Ensemble application of convolutional and recurrent neural networks for multi-label text categorization. In: International Joint Conference on Neural Networks (IJCNN). New York: IEEE, 2017. 2377--2383. Google Scholar

[14] Zhang W, Yan J, Wang X, et al. Deep Extreme Multi-label Learning. In: Proceedings of the 2018 ACM on International Conference on Multimedia Retrieval. New York: ACM, 2018. 100--107. Google Scholar

[15] Wright J, Yang A Y, Ganesh A. Robust face recognition via sparse representation.. IEEE Trans Pattern Anal Mach Intell, 2009, 31: 210-227 CrossRef PubMed Google Scholar

[16] Yang M, Zhang L, Yang J, et al. Metaface learning for sparse representation based face recognition. In: Proceedings of IEEE International Conference on Image Processing. New York: IEEE, 2010. 1601--1604. Google Scholar

[17] Zhang Q, Li B. Discriminative K-SVD for dictionary learning in face recognition. In: Proceedings of Computer Vision and Pattern Recognition. New York: IEEE, 2010. 2691--2698. Google Scholar

[18] Ramirez I, Sprechmann P, Sapiro G. Classification and clustering via dictionary learning with structured incoherence and shared features. In: Proceedings of Computer Vision and Pattern Recognition. New York: IEEE, 2010. 3501--3508. Google Scholar

[19] Mairal J, Bach F, Ponce J, et al. Discriminative learned dictionaries for local image analysis. In: Proceedings of IEEE Conference on Computer Vision and Pattern Recognition. New York: IEEE, 2008. 1--8. Google Scholar

[20] Peng J. Learning inter-related visual dictionary for object recognition. In: Proceedings of Computer Vision and Pattern Recognition. New York: IEEE, 2012. 3490--3497. Google Scholar

[21] Kong S, Wang D. A dictionary learning approach for classification: separating the particularity and the commonality. In: Proceedings of European Conference on Computer Vision. Berlin: Springer-Verlag, 2012. 186--199. Google Scholar

[22] Gao S H, Tsang I W H, Ma Y. Learning Category-Specific Dictionary and Shared Dictionary for Fine-Grained Image Categorization. IEEE Trans Image Process, 2014, 23: 623-634 CrossRef PubMed ADS Google Scholar

[23] Gu S, Zhang L, Zuo W, et al. Projective dictionary pair learning for pattern classification. In: Proceedings of Advances in Neural Information Processing Systems. 2014. 793--801. Google Scholar

[24] Yang M, Zhang L, Feng X. Sparse Representation Based Fisher Discrimination Dictionary Learning for Image Classification. Int J Comput Vis, 2014, 109: 209-232 CrossRef Google Scholar

[25] Lu J, Wang G, Zhou J. Simultaneous Feature and Dictionary Learning for Image Set Based Face Recognition. IEEE Trans Image Process, 2017, 26: 4042-4054 CrossRef PubMed ADS Google Scholar

[26] Pham D S, Venkatesh S. Joint learning and dictionary construction for pattern recognition. In: Proceedings of IEEE Conference on Computer Vision and Pattern Recognition. New York: IEEE, 2008. 1--8. Google Scholar

[27] Mairal J, Ponce J, Sapiro G, et al. Supervised dictionary learning. In: Proceedings of Advances in neural information processing systems, 2009. 1033--1040. Google Scholar

[28] Zhuolin Jiang , Zhe Lin , Davis L S. Label Consistent K-SVD: Learning a Discriminative Dictionary for Recognition. IEEE Trans Pattern Anal Mach Intell, 2013, 35: 2651-2664 CrossRef Google Scholar

[29] Cao X C, Zhang H, Guo X J. SLED: Semantic Label Embedding Dictionary Representation for Multilabel Image Annotation. IEEE Trans Image Process, 2015, 24: 2746-2759 CrossRef PubMed ADS Google Scholar

[30] Jing X Y, Wu F, Li Z. Multi-Label Dictionary Learning for Image Annotation. IEEE Trans Image Process, 2016, 25: 2712-2725 CrossRef PubMed ADS Google Scholar

[31] Zhao Q, Li Q M. Extraction and application of Yi nationality's decorative patterns in the tourist souvenirs. Packag Eng, 2017, 38: 240--244. Google Scholar

[32] Wang Y. The digital applications of folk art elements. Taiyuan: Taiyuan University of Technology, 2014. Google Scholar

[33] He N, Lu K. An Image Segmentation Method for Chinese Paintings by Combining Deformable Models with Graph Cuts. In: Proceedings of the 14th International Conference on Human-Computer Interaction: Design and Development Approaches. Berlin: Springer, 2011. 571--579. Google Scholar

[34] Xiang J, Zhang J, Pan R R, et al. Graphic contour extraction for printed fabric based on texture smoothing. J Textile Res, 2017, 38: 162--167. Google Scholar

[35] Yang B W, Lu X Q, Chen C. Multiscale color segmentation algorithm for textile images. J Zhejiang Univ (Eng Sci), 2005, 39: 530--533+546. Google Scholar

[36] Yue J F. On the subject classification and cultural connotation of paper-cut in Bozhou. Home Darma, 2017, 10: 255--256. Google Scholar

[37] Yang J L. A study on classification of traditional tibetan clothing patterns. Big Stage, 2015, 325: 230--231. Google Scholar

[38] Liu C H. Classification and artistic characteristics of mongolian traditional patterns. Big Stage, 2015, 3: 81--82. Google Scholar

[39] Zhang H. Research and application of image classification of stylized fabrics. Dissertation for Master Degree. Zhejiang: Zhejiang University, 2012. Google Scholar

[40] Qian J Q. A study on semantic-based thangka image annotation and retrieval. Lanzhou: Northwest Minzu University, 2010. Google Scholar

[41] Zhang D Q, Pham B, Li Y F. Modelling traditional Chinese paintings for content-based image classification and retrieval. In: Proceedings of the 10th International Multimedia Modelling Conference. Washington: IEEE Computer Society, 2004. 258. Google Scholar

[42] Pawening R E, Dijaya R, Brian T, et al. Classification of textile image using support vector machine with textural feature. In: Proceedings of International Conference on Information & Communication Technology and Systems. New York: IEEE, 2016. 119--122. Google Scholar

[43] An X L, Wei K Y. The discuss of the application of Jingpo ethnic traditional national pattern in landscape architecture. South China Agr (Garden Flower Edition) 2011, 5: 1--6. Google Scholar

[44] Wang Y. A study on application of Uygur national patterns in food packaging design. Dissertation for Master Degree. Xi'an: Northwest University, 2015. Google Scholar

[45] Zhao H Y, Peng H, Zhang X L. A generation method of carpet pattern based on evolutionary computation. J Graph, 2015, 36: 41--46. Google Scholar

[46] Zhao H Y, Pan Z G, Xu Z G. Automatic generation of Xinjiang ethnic fabric patterns based on configuration style. J Graph, 2013, 34: 17--21. Google Scholar

[47] Zhao H Y, Yang Y F, Xu G M. Design method for xinjiang folk art pattern. Comput Syst Appl, 2011, 20: 94--99. Google Scholar

[48] Zhao H Y, Peng H, Yang Y F, et al. A generation method of carpet pattern based on topology configuration. J Comput-Aided Design Comput Graph, 2013, 25: 502--509. Google Scholar

[49] Zhang M L, Wu L. Lift: multi-label learning with label-specific features. IEEE Transactions on Pattern Analysis & Machine Intelligence, 2015, 37:107-120 DOI: 10.5591/978-1-57735-516-8/IJCAI11-270. Google Scholar

[50] Wang J J, Yang J C, Yu K, et al. Locality-constrained linear coding for image classification. In: Proceedings of IEEE Computer Society Conference on Computer Vision and Pattern Recognition, San Francisco, 2010. 3360--3367. Google Scholar

[51] Zhang M L, Pe?a J M, Robles V. Feature selection for multi-label naive Bayes classification. Inf Sci, 2009, 179: 3218-3229 CrossRef Google Scholar

[52] Huang J, Li G R, Huang Q M, et al. Learning label specific features for multi-label classification. In: Proceedings of IEEE International Conference on Data Mining, Atlantic City, 2015. 181--190. Google Scholar

  • Figure 1

    (Color online) Roc pattern in miao nationality. (a) Body of the fish; (b) wings of the bird; (c) body of the bird

  • Figure 2

    Framework of national cultural pattern digitalization

  • Figure 3

    (Color online) Research examples of national cultural pattern digitalization

  • Figure 4

    (Color online) Framework of multi-label image annotation. (a) Represents image feature extraction; (b) constructs the dictionary; (c) and (d) are the test stages

  • Figure 5

    (Color online) Examples of image annotation. (a) Traditional carpet pattern; (b) traditional national dress pattern; (c) Ming and Qing court dress pattern

  • Figure 6

    Convergence of the objective function

  • Figure 7

    (Color online) Hyper-parameters: (a) $\lambda_1$; (b) $\lambda_2$; (c) $\lambda_3$; (d) $\gamma$

  • Table 1   Attributions of national cultural pattern data sets
    Data set Number of Number of LCard LDen DL PDL Number of trianing
    images labels samples
    Ming and Qing court dress pattern 899 5 1.3582 0.2716 12 0.0133 600
    Traditional national dress pattern 782 6 2.7724 0.4621 18 0.0230 400
    Traditional carpet pattern 536 3 1.5690 0.5230 7 0.0131 200
  • Table 2   Experimental results of Ming and Qing court dress pattern data sets
    One-error Coverage Ranking-loss Average-precision
    ML-KNN 0.3946 1.1873 0.2152 0.7419
    Rank-SVM 0.3679 1.1037 0.1909 0.7653
    MLNB 0.5552 1.5418 0.3088 0.6533
    LLSF-BR 0.4013 1.2642 0.2305 0.7340
    LLSF-CC 0.7458 1.8428 0.3946 0.5427
    LLSF 0.5686 1.6923 0.3464 0.6243
    SCMIDL 0.3478 0.9565 0.1577 0.7911
  • Table 3   Experimental results of traditional national dress pattern data sets
    One-error Coverage Ranking-loss Average-precision
    ML-KNN 0.0700 1.7853 0.0586 0.9313
    Rank-SVM 0.0497 1.7382 0.0550 0.9431
    MLNB 0.0497 1.8508 0.0699 0.9321
    LLSF-BR 0.0445 1.7304 0.0528 0.9459
    LLSF-CC 0.0602 1.7592 0.0620 0.9385
    LLSF 0.0524 1.8455 0.0667 0.9357
    SCMIDL 0.0445 1.6780 0.0450 0.9524
  • Table 4   Experimental results of traditional carpet pattern data sets
    One-error Coverage Ranking-loss Average-precision
    ML-KNN 0.1380 0.8661 0.1887 0.9049
    Rank-SVM 0.0675 0.7589 0.1196 0.9445
    MLNB 0.2791 1.0327 0.3037 0.8341
    LLSF-BR 0.0859 0.7827 0.1365 0.9351
    LLSF-CC 0.0951 0.8393 0.1672 0.9225
    LLSF 0.0767 0.7173 0.1043 0.9496
    SCMIDL 0.0670 0.7366 0.0950 0.9530
  • Table 5   Experimental results of different objective function on data set 1/2/3
    Data set Original objective function $\rm~\lambda_{2}=0$, $\rm~\lambda_{3}=0$, $\rm~\gamma=0$ $\rm~\lambda_{2}=0$ $\rm~\lambda_{3}=0$, $\rm~\gamma=0$
    3*Average-precision 1 0.7865 0.7829 0.7862 0.7809
    2 0.9459 0.9411 0.9467 0.9424
    3 0.9305 0.9277 0.9290 0.9277
    3*Coverage 1 1.0736 1.0803 1.0702 1.0870
    2 1.6571 1.6518 1.6492 1.6466
    3 0.8184 0.8244 0.8214 0.8244
    3*One-error 1 0.3378 0.3445 0.3411 0.3478
    2 0.0340 0.0628 0.0340 0.0602
    3 0.1018 0.1080 0.1049 0.1080
    3*Ranking-loss 1 0.1826 0.1856 0.1826 0.1873
    2 0.0509 0.0525 0.0499 0.0516
    3 0.1388 0.1435 0.1404 0.1435

Copyright 2019 Science China Press Co., Ltd. 《中国科学》杂志社有限责任公司 版权所有

京ICP备18024590号-1