Abstract
Structural models are shown to be highly effective tools in computer vision and image processing. Conditional random fields (CRFs) are a powerful group of statistical graphical models in the field of structured prediction. Different learning methods are used to train the CRFs. In this paper, a novel class of losses is proposed for the purpose of CRF learning. The proposed losses are categorized into marginal-based loss groups which have a different view comparing to likelihood-based training losses. The proposed losses are compared with existing losses and their performances are evaluated through three experiments. The results indicate superior efficiency of the proposed framework.
Similar content being viewed by others
References
Ahmadi, E., Azimifar, Z., Fieguth, P., Ayatollahi, S.: Discriminative graphical model for porous media image synthesis. Iran. J. Sci. Technol. 38, 177–189 (2014)
Besag, J.: Statistical analysis of non-lattice data. Stat. 24, 179–195 (1975)
Cramer, H.: Mathematical Methods of Statistics. Princeton University Press, Princeton (1999)
Cristianini, N., Shawe-Taylor, J.: An Introduction to Support Vector Machines and Other Kernel-Based Learning Methods. Cambridge University Press, Cambridge (2000)
Dalal, N., Triggs, B.: Histograms of oriented gradients for human detection. In: IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR), vol. 1, pp. 886–893 (2005)
Domke, J.: Learning graphical model parameters with approximate marginal inference. IEEE Trans. Pattern Anal. Mach. Intell. 35(10), 2454–2467 (2013)
Hazan, T., Shashua, A.: Norm-product belief propagation: primal-dual message-passing for approximate inference. IEEE Trans. Inform. Theory 56(12), 6294–6316 (2010)
Jordan, A.: On discriminative vs. generative classifiers: a comparison of logistic regression and naive Bayes. Adv. Neural Inform. Process. Syst. 14, 841 (2002)
Kumar, S., Hebert, M.: Discriminative fields for modeling spatial dependencies in natural images. NIPS (2003)
Lafferty, J., McCallum, A., Pereira, F.C.: Conditional random fields: probabilistic models for segmenting and labeling sequence data. In: ICML, pp. 282–289 (2001)
Levin, A., Weiss, Y.: Learning to combine bottom-up and top-down segmentation. Int. J. Comput. Vis. 81(1), 105–118 (2009)
Marroquin, J., Mitter, S., Poggio, T.: Probabilistic solution of ill-posed problems in computational vision. J. Am. Stat. Assoc. 82(397), 76–89 (1987)
Masnadi-Shirazi, H., Vasconcelos, N.: On the design of loss functions for classification: theory, robustness to outliers, and savageboost. In: Advances in Neural Information Processing Systems, pp. 1049–1056 (2009)
Masnadi-Shirazi, H., Vasconcelos, N.: Variable margin losses for classifier design. In: Advances in Neural Information Processing Systems, pp. 1576–1584 (2010)
Nowozin, S., Gehler, P.V., Lampert, C.H.: On parameter learning in crf-based approaches to object class image segmentation. In: Computer Vision–ECCV 2010, Springer, pp. 98–111 (2010)
Nowozin, S., Lampert, C.H.: Structured learning and prediction in computer vision. Founda. Trends Comput. Graph. Vis. 6(3–4), 185–365 (2011)
PVS, A., Karthik, G.: Part-of-speech tagging and chunking using conditional random fields and transformation based learning. Shallow Parsing for South Asian Languages 21 (2007)
Savage, L.J.: Elicitation of personal probabilities and expectations. J. Am. Stat. Assoc. 66(336), 783–801 (1971)
Shafiee, M., Azimifar, Z., Fieguth, P.: Model-based tracking: Temporal conditional random fields. In: 2010 17th IEEE International Conference on Image Processing (ICIP), pp. 4645–4648 (2010)
Sutton, C., McCallum, A.: An introduction to conditional random fields. Mach. Learn. 4(4), 267–373 (2011)
Torralba, A., Murphy, K.P., Freeman, W.T.: Contextual models for object detection using boosted random fields. Adv. Neural Inform. Process. Syst. 17, 1401–1408 (2004)
Toyoda, T., Hasegawa, O.: Random field model for integration of local information and global information. IEEE Trans. Pattern Anal. Mach. Intell. 30(8), 1483–1489 (2008)
Vishwanathan, S., Schraudolph, N.N., Schmidt, M.W., Murphy, K.P.: Accelerated training of conditional random fields with stochastic gradient methods. In: Proceedings of the 23rd International Conference on Machine Learning, pp. 969–976, ACM (2006)
Wainwright, M .J., J, T.S., Willsky, A.S.: A new class of upper bounds on the log partition function. IEEE Trans. Inform. Theory 51(7), 2313–2335 (2005)
Wainwright, M.J., Jordan, M.I.: Graphical models, exponential families, and variational inference. Found. Trends Mach. Learn. 1(1–2), 1–305 (2008)
Wang, C., Komodakis, N., Paragios, N.: Markov random field modeling, inference & learning in computer vision & image understanding: a survey. Comput. Vis. Image Underst. 117(11), 1610–1627 (2013)
Wang, S., Xu, R., Liu, B., Gui, L., Zhou, Y.: Financial named entity recognition based on conditional random fields and information entropy. In: 2014 International Conference on Machine Learning and Cybernetics (ICMLC), vol. 2, pp. 838–843 (2014)
Yu, D., Deng, L.: Hidden Markov models and the variants. In: Automatic Speech Recognition, Springer, pp. 23–54 (2015)
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
About this article
Cite this article
Ahmadi, E., Azimifar, Z. Margin Losses for Training Conditional Random Fields. J Math Imaging Vis 56, 499–510 (2016). https://doi.org/10.1007/s10851-016-0651-y
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s10851-016-0651-y