Neural Computing and Applications

, Volume 29, Issue 10, pp 733–743 | Cite as

Stochastic learning of multi-instance dictionary for earth mover’s distance-based histogram comparison

  • Jihong Fan
  • Ru-Ze Liang
Original Article


Dictionary plays an important role in multi-instance data representation. It maps bags of instances to histograms. Earth mover’s distance (EMD) is the most effective histogram distance metric for the application of multi-instance retrieval. However, up to now, there is no existing multi-instance dictionary learning methods designed for EMD-based histogram comparison. To fill this gap, we develop the first EMD-optimal dictionary learning method using stochastic optimization method. In the stochastic learning framework, we have one triplet of bags, including one basic bag, one positive bag, and one negative bag. These bags are mapped to histograms using a multi-instance dictionary. We argue that the EMD between the basic histogram and the positive histogram should be smaller than that between the basic histogram and the negative histogram. Base on this condition, we design a hinge loss. By minimizing this hinge loss and some regularization terms of the dictionary, we update the dictionary instances. The experiments over multi-instance retrieval applications shows its effectiveness when compared to other dictionary learning methods over the problems of medical image retrieval and natural language relation classification.


Multi-instance learning Multi-instance dictionary Histogram comparision Earth mover’s distance Stochastic learning Medical image retrieval 



The work was funded by Science and Technology project under Grant No. 12531826 of Education Department, Heilongjiang, China.


  1. 1.
    Beecks C, Uysal M, Seidl T (2016) Earth mover’s distance vs. quadratic form distance: an analytical and empirical comparison. In: Proceedings—2015 IEEE international symposium on multimedia, ISM 2015, pp 233–236Google Scholar
  2. 2.
    Chen Y, Bi J, Wang JZ (2006) Miles: multiple-instance learning via embedded instance selection. IEEE Trans Pattern Anal Mach Intell 28(12):1931–1947CrossRefGoogle Scholar
  3. 3.
    Chen YH, Chen TC, Ma TC, Lee TH, Chen LG, et al (2009) Sub-microwatt knn classifier for implantable closed-loop epileptic neuromodulation system. In: Proceedings of the 2009 international symposium on bioelectronics and bioinformatics. RMIT University, School of Electrical and Computer Engineering, p 13Google Scholar
  4. 4.
    Chen TC, Chen YY, Ma TC, Chen LG (2011) Design and implementation of cubic spline interpolation for spike sorting microsystems. In: 2011 IEEE international conference on acoustics, speech and signal processing (ICASSP). IEEE, pp 1641–1644Google Scholar
  5. 5.
    Chen TC, Ma TC, Chen YY, Chen LG (2012) Low power and high accuracy spike sorting microprocessor with on-line interpolation and re-alignment in 90 nm cmos process. In: 2012 annual international conference of the IEEE engineering in medicine and biology society. IEEE, pp 4485–4488Google Scholar
  6. 6.
    Chen TT, Liu C, Ding XM, Zou H, Shen Q, Liu Y (2016) A multi-instance multi-label learning algorithm based on feature selection. In: Proceedings—2015 10th international conference on broadband and wireless computing, communication and applications, BWCCA 2015, pp 587–590Google Scholar
  7. 7.
    Clarkson E, Cushing J (2016) Shannon information and receiver operating characteristic analysis for multiclass classification in imaging. J Opt Soc Am A Opt Image Sci Vis 33(5):930–937CrossRefGoogle Scholar
  8. 8.
    Fan X, Tang K (2010) Enhanced maximum AUC linear classifier. In: 2010 seventh international conference on fuzzy systems and knowledge discovery (FSKD), vol 4. IEEE, pp 1540–1544Google Scholar
  9. 9.
    Fan X, Yuan C (2015) An improved lower bound for Bayesian network structure learning. In: AAAI, pp 3526–3532Google Scholar
  10. 10.
    Fan X, Malone B, Yuan C (2014) Finding optimal bayesian network structures with constraints learned from data. In: Proceedings of the 30th annual conference on uncertainty in artificial intelligence (UAI-14), pp 200–209Google Scholar
  11. 11.
    Fan X, Yuan C, Malone BM (2014) Tightening bounds for bayesian network structure learning. In: AAAI. Citeseer, pp 2439–2445Google Scholar
  12. 12.
    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
  13. 13.
    Goadrich M, Oliphant L, Shavlik J (2006) Gleaner: creating ensembles of first-order clauses to improve recall–precision curves. Mach Learn 64(1–3):231–261CrossRefzbMATHGoogle Scholar
  14. 14.
    Hendrickx, I, Kim SN, Kozareva Z, Nakov P, Ó Séaghdha D, Padó S, Pennacchiotti M, Romano L, Szpakowicz S (2009) Semeval-2010 task 8: multi-way classification of semantic relations between pairs of nominals. In: Proceedings of the workshop on semantic evaluations: recent achievements and future directions. Association for Computational Linguistics, pp 94–99Google Scholar
  15. 15.
    Huo J, Gao Y, Yang W, Yin H (2012) Abnormal event detection via multi-instance dictionary learning. Lecture notes in computer science (including subseries lecture notes in artificial intelligence and lecture notes in bioinformatics) 7435 LNCS, pp 76–83Google Scholar
  16. 16.
    Huo J, Gao Y, Yang W, Yin H (2014) Multi-instance dictionary learning for detecting abnormal events in surveillance videos. Int J Neural Syst 24(3):1430,010CrossRefGoogle Scholar
  17. 17.
    Ju CC, Liu TM, Chen YH, Lee KB, Cheng CY, Chao HT, Wang CM, Wu TH, Lin TA, Chou HL, et al (2012) A 1.94 mm 2, 38.17 mw dual vp8/h. 264 full-hd encoder/decoder lsi for social network services (sns) over smart-phones. In: 2012 IEEE Asian solid state circuits conference (A-SSCC). IEEE, pp 13–16Google Scholar
  18. 18.
    Kim M, Han D, Ko H (2016) Joint patch clustering-based dictionary learning for multimodal image fusion. Inf Fusion 27:198–214CrossRefGoogle Scholar
  19. 19.
    King DR, Li W, Squiers JJ, Mohan R, Sellke E, Mo W, Zhang X, Fan W, DiMaio JM, Thatcher JE (2015) Surgical wound debridement sequentially characterized in a porcine burn model with multispectral imaging. Burns 41(7):1478–1487CrossRefGoogle Scholar
  20. 20.
    Li JY, Li JH, Shui-Cheng Y (2015) Multi-instance learning using information entropy theory for image retrieval. In: Proceedings—17th IEEE international conference on computational science and engineering, CSE 2014, jointly with 13th IEEE international conference on ubiquitous computing and communications, IUCC 2014, 13th international symposium on pervasive systems, algorithms, and networks, I-SPAN 2014 and 8th international conference on frontier of computer science and technology, FCST 2014, pp 1727–1733Google Scholar
  21. 21.
    Li W, Mo W, Zhang X, Lu Y, Squiers JJ, Sellke EW, Fan W, DiMaio JM, Thatcher JE (2015) Burn injury diagnostic imaging device’s accuracy improved by outlier detection and removal. In: SPIE Defense + Security. International Society for Optics and Photonics, pp 947206–947206Google Scholar
  22. 22.
    Li W, Mo W, Zhang X, Squiers JJ, Lu Y, Sellke EW, Fan W, DiMaio JM, Thatcher JE (2015) Outlier detection and removal improves accuracy of machine learning approach to multispectral burn diagnostic imaging. J Biomed Opt 20(12):121305–121305CrossRefGoogle Scholar
  23. 23.
    Liang RZ, Shi L, Wang H, Meng J, Wang JJY, Sun Q, Gu Y (2016) Optimizing top precision performance measure of content-based image retrieval by learning similarity function. In: 2016 23st international conference on pattern recognition (ICPR). IEEEGoogle Scholar
  24. 24.
    Lin F, Wang J, Zhang N, Xiahou J, McDonald N (2016) Multi-kernel learning for multivariate performance measures optimization. Neural Comput Appl 1–13Google Scholar
  25. 25.
    Liu X, Wang J, Yin M, Edwards B, Xu P (2015) Supervised learning of sparse context reconstruction coefficients for data representation and classification. Neural Comput Appl 1–9Google Scholar
  26. 26.
    Lobel H, Vidal R, Soto A (2015) Learning shared, discriminative, and compact representations for visual recognition. IEEE Trans Pattern Anal Mach Intell 37(11):2218–2231CrossRefGoogle Scholar
  27. 27.
    Ma TC, Chen TC, Chen LG (2014) Design and implementation of a low power spike detection processor for 128-channel spike sorting microsystem. In: 2014 IEEE international conference on acoustics, speech and signal processing (ICASSP). IEEE, pp 3889–3892Google Scholar
  28. 28.
    Mo W, Mohan R, Li W, Zhang X, Sellke EW, Fan W, DiMaio JM, Thatcher JE (2015) The importance of illumination in a non-contact photoplethysmography imaging system for burn wound assessment. In: SPIE BiOS. International Society for Optics and Photonics, pp 93030M–93030MGoogle Scholar
  29. 29.
    Nabavi S, Beck A (2016) Earth mover’s distance for differential analysis of heterogeneous genomics data. In: 2015 IEEE global conference on signal and information processing, GlobalSIP 2015, pp 963–966Google Scholar
  30. 30.
    Rose C, Turi D, Williams A, Wolstencroft K, Taylor C (2006) Web services for the DDSM and digital mammography research. Lecture notes in computer science (including subseries lecture notes in artificial intelligence and lecture notes in bioinformatics) 4046 LNCS, pp 376–383Google Scholar
  31. 31.
    Shao Z, Er M (2016) Efficient leave-one-out cross-validation-based regularized extreme learning machine. Neurocomputing 194:260–270CrossRefGoogle Scholar
  32. 32.
    Tong J, Schreier P, Guo Q, Tong S, Xi J, Yu Y (2016) Shrinkage of covariance matrices for linear signal estimation using cross-validation. IEEE Trans Signal Process 64(11):2965–2975MathSciNetCrossRefGoogle Scholar
  33. 33.
    Vanwinckelen G, Tragante do OV, Fierens D, Blockeel H (2016) Instance-level accuracy versus bag-level accuracy in multi-instance learning. Data Min Knowl Disc 30(2):313–341MathSciNetCrossRefGoogle Scholar
  34. 34.
    Wang X, Kambhamettu C (2013) Gender classification of depth images based on shape and texture analysis. In: 2013 IEEE global conference on signal and information processing (GlobalSIP). IEEE, pp 1077–1080Google Scholar
  35. 35.
    Wang X, Kambhamettu C (2014) Leveraging appearance and geometry for kinship verification. In: 2014 IEEE international conference on image processing (ICIP). IEEE, pp 5017–5021Google Scholar
  36. 36.
    Wang H, Wang J (2014) An effective image representation method using kernel classification. In: 2014 IEEE 26th international conference on tools with artificial intelligence (ICTAI 2014), pp 853–858Google Scholar
  37. 37.
    Wang X, Ly V, Lu G, Kambhamettu C (2013) Can we minimize the influence due to gender and race in age estimation? In: 2013 12th international conference on machine learning and applications (ICMLA). IEEE, vol 2, pp 309–314Google Scholar
  38. 38.
    Wang X, Wang B, Bai X, Liu W, Tu Z (2013) Max-margin multiple-instance dictionary learning. In: Proceedings of the 30th international conference on machine learning, pp 846–854Google Scholar
  39. 39.
    Wang J, Wang H, Zhou Y, McDonald N (2015) Multiple kernel multivariate performance learning using cutting plane algorithm. In: 2015 IEEE international conference on systems, man, and cybernetics (SMC). IEEE, pp 1870–1875Google Scholar
  40. 40.
    Wang J, Zhou Y, Duan K, Wang JJY, Bensmail H (2015) Supervised cross-modal factor analysis for multiple modal data classification. In: 2015 IEEE international conference on systems, man, and cybernetics (SMC). IEEE, pp 1882–1888Google Scholar
  41. 41.
    Wang X, Guo R, Kambhamettu C (2015) Deeply-learned feature for age estimation. In: 2015 IEEE winter conference on applications of computer vision. IEEE, pp 534–541Google Scholar
  42. 42.
    Wang D, Attwood K, Tian L (2016) Receiver operating characteristic analysis under tree orderings of disease classes. Stat Med 35(11):1907–1926MathSciNetCrossRefGoogle Scholar
  43. 43.
    Wang K, Liu J, González D (2016) Domain transfer multi-instance dictionary learning. Neural Comput Appl 1–10Google Scholar
  44. 44.
    Wu JH, Tang LB, Zhao BJ, Deng CW, Li JT (2015) Visual dictionary and online multi-instance learning based object tracking. Xi Tong Gong Cheng Yu Dian Zi Ji Shu/Syst Eng Electr 37(2):428–435Google Scholar
  45. 45.
    Xia H, Hoi S, Jin R, Zhao P (2014) Online multiple kernel similarity learning for visual search. IEEE Trans Pattern Anal Mach Intell 36(3):536–549CrossRefGoogle Scholar
  46. 46.
    Xu S, Lu W, Xu L (2012) Push-and pull-based epidemic spreading in networks: thresholds and deeper insights. ACM Trans Auton Adapt Syst (TAAS) 7(3):32Google Scholar
  47. 47.
    Xu L, Zhan Z, Xu S, Ye K (2013) Cross-layer detection of malicious websites. In: Proceedings of the third ACM conference on data and application security and privacy. ACM, pp 141–152Google Scholar
  48. 48.
    Xu L, Zhan Z, Xu S, Ye K (2014) An evasion and counter-evasion study in malicious websites detection. In: 2014 IEEE conference on communications and network security (CNS). IEEE, pp 265–273Google Scholar
  49. 49.
    Xu S, Lu W, Xu L, Zhan Z (2014) Adaptive epidemic dynamics in networks: thresholds and control. ACM Trans Auton Adapt Syst (TAAS) 8(4):19Google Scholar
  50. 50.
    Xu M, Dong H, Chen C, Li L (2016) Unsupervised dictionary learning with fisher discriminant for clustering. Neurocomputing 194:65–73CrossRefGoogle Scholar
  51. 51.
    Yan Z, Zhan Y, Peng Z, Liao S, Shinagawa Y, Zhang S, Metaxas D, Zhou X (2016) Multi-instance deep learning: discover discriminative local anatomies for bodypart recognition. IEEE Trans Med Imaging 35(5):1332–1343CrossRefGoogle Scholar
  52. 52.
    Ying P, Liu J, Lu H (2015) Dictionary learning based superpixels clustering for weakly-supervised semantic segmentation. In: Proceedings—international conference on image processing, ICIP, vol 2015-December, pp 4258–4262Google Scholar
  53. 53.
    Zhang P, Su W (2012) Statistical inference on recall, precision and average precision under random selection. In: Proceedings—2012 9th international conference on fuzzy systems and knowledge discovery, FSKD 2012, pp 1348–1352Google Scholar
  54. 54.
    Zhang M, Peng J, Liu X (2016) Sparse coding with Earth movers distance for multi-instance histogram representation. Neural Comput Appl 1–12Google Scholar

Copyright information

© The Natural Computing Applications Forum 2016

Authors and Affiliations

  1. 1.Qiqihar Medical UniversityQiqiharPeople’s Republic of China
  2. 2.King Abdullah University of Science and TechnologyThuwalSaudi Arabia

Personalised recommendations