Presentation on theme: "Automatic Image Annotation Using Group Sparsity"— Presentation transcript:
1 Automatic Image Annotation Using Group Sparsity Shaoting Zhang1, Junzhou Huang1,Yuchi Huang1, Yang Yu1, Hongsheng Li2,Dimitris Metaxas11CBIM, Rutgers University, NJ2IDEA Lab, Lehigh University, PA
2 IntroductionsGoal: image annotation is to automatically assign relevant text keywords to any given image, reflecting its content.Previous methods:Topic models [Barnard, et.al., J. Mach. Learn Res.’03; Putthividhya, et.al., CVPR’10]Mixture models [Carneiro, et.al., TPAMI’07; Feng, et.al., CVPR’04]Discriminative models [Grangier, et.al., TPAMI’08; Hertz, et.al., CVPR’04]Nearest neighbor based methods [Makadia, et.al., ECCV’08; Guillaumin, et.al., ICCV’09]Add more references
3 Introductions Limitations: Our method and contributions: Features are often preselected, yet the properties of different features and feature combinations are not well investigated in the image annotation task.Feature selection is not well investigated in this application.Our method and contributions:Use feature selection to solve annotation problem.Use clustering prior and sparsity prior to guide the selection.
4 Outline Regularization based Feature Selection Obtain Image Pairs Annotation frameworkL2 norm regularizationL1 norm regularizationGroup sparsity based regularizationObtain Image PairsExperiments
5 Regularization based Feature Selection Given similar/dissimilar image pair list (P1,P2)………………Note that we use absolute value for the difference.FP1FP2X
7 Regularization based Feature Selection Annotation frameworkWeightsSimilarityTestinginputHighsimilarityTraining data
8 Regularization based Feature Selection L2 regularizationRobust, solvable: (XTX+λI)-1XTYNo sparsity%L2 norm tries to produce small weights. However, usually it cannot push weights to zero. The intuitive explanation is that the magnitude of the slope of a quadratic function decreases when approaching zero (magnitude of slope will linearly decrease to zero when approaching zero). Thus the penalty assigned to the weight changing also decreases. Large weight has large penalty. Thus it’s not preferred. However, small weight has almost no difference with zero weight. Thus there is generally no penalty.wHistogram of weights
9 Regularization based Feature Selection L1 regularizationConvex optimizationBasis pursuit, Grafting, Shooting, etc.Sparsity prior%In this case the magnitude of slope is constant (except for 0, which is not differentiable). Thus the weights will be pushed constantly towards zero. Furthermore, it’s not so sensitive for large weights compared to L2 norm.wHistogram of weights
10 Regularization based Feature Selection RGBHSVGroup sparsityL2 inside the same group, L1 for different groupsBenefits: removal of whole feature groupsProjected-gradientFirst we need to divide groups manually. In this case, we just naturally define RGB, HSV, etc. as different groups.Within the same group, we use L2. For different groups, we use L1.The intuition is that we either push the whole group to zero, or keep the whole group small (but nonzero).=0≠0 M. Yuan and Y. Lin. Model selection and estimation in regression with grouped variables. Journal of the Royal Statistical Society, Series B, 68:49–67, 2006. E. Berg, M. Schmidt, M. Friedlander, and K. Murphy. Group sparsity via linear-time projection. In Technical report, TR ,
11 Outline Regularization based Feature Selection Obtain Image Pairs Only rely on keyword similarityAlso rely on feedback informationExperiments
12 Obtain Image PairsPrevious method solely relies on keyword similarity, which induces a lot of noise.Traditional method assumes that images sharing more than 3 keywords are similar, and images having no common keyword are dissimilar.However, similar keywords do not necessary mean that their feature distances are close. In this case (left figure), although most pairs have small distance in feature space, there are still a lot of exceptions. Combine both similar and dissimilar pairs together, it is difficult to linearly separate them using distance measurement.Furthermore, using this method, the number of dissimilar images is much larger than the one of similar images, which will bias the training.Distance histogram of similar pairsDistance histogram of all pairs A. Makadia, V. Pavlovic, and S. Kumar. A new baseline for image annotation. In ECCV, pages 316–329, 2008.
13 Obtain Image PairsInspired by the relevance feedback and the expectation maximization method.k1 nearestk2 farthest(candidates ofsimilar pairs)(candidates ofdissimilar pairs)Using our method, the noises of similar image pairs (positive sample) are much less.
14 Outline Regularization based Feature Selection Obtain Image Pairs ExperimentsExperimental settingsEvaluation of regularization methodsEvaluation of generalitySome annotation results
15 Experimental Settings Data protocolsCorel5K (5k images)IAPR TC12 (20k images)EvaluationAverage precisionAverage recall#keywords recalled (N+) M. Grubinger, P. D. Clough, H. Muller, and T. Deselaers. The iapr tc-12 benchmark - a new evaluation resource for visual information systems
16 Experimental Settings FeaturesRGB, HSV, LABOpponentrghistogramTransformed color distributionColor from SaliencyHaar, GaborSIFT, HOG X. Hou and L. Zhang. Saliency detection: A spectral residual approach. In CVPR, 2007. A. Makadia, V. Pavlovic, and S. Kumar. A new baseline for image annotation. In ECCV, pages 316–329, 2008. K. van de Sande, T. Gevers, and C. Snoek. Evaluating color descriptors for object and scene recognition. PAMI, 99(1),2010. N. Dalal and B. Triggs. Histograms of oriented gradients for human detection. In CVPR, pages 886–893, 2005.
17 Evaluation of Regularization Methods PrecisionRecallN+Corel5KIIAPR TC12
18 Evaluation of Generality Weights computed from Corel5K, then applied on IAPR TC12.PrecisionRecallN+λλλ
19 Some Annotation Results Since we transfer 5 keywords every time (while the ground truth may only have 2-4 keywords), our precision is adversely affected. There may be redundancy in predicted keywords.However, as we will see, some keywords (not in ground truth) actually describe the image well. In other words, they are better than human annotation in some sense.
20 Conclusions and Future Work Proposed a feature selection framework using both sparsity and clustering priors to annotate images.The sparse solution improves the scalability.Image pairs from relevance feedback perform much better.Future workDifferent grouping methods.Automatically find groups (dynamic group sparsity).More priors (combine with other methods).Extend this framework to object recognition.