Abstract
In this paper we propose an information-geometric method for comparing superpixel (turbopixel) images. Turbopixels are encoded by tensors and they are referred to as TurboTensors. Our methodology has three ingredients. Firstly, we formulate the comparison of the turbopixels topology in terms of the non-rigid alignment of the Isomap embedding of the weighted adjacency matrices; we propose a multi-dimensional information-theoretic dissimilarity measure. Secondly, we formulate the comparison of bags-of-turbopixels through tangent spaces de-projection and multi-dimensional and non-parametric information-theoretic dissimilarity measures. Thirdly, we combine the two latter elements into a flexible energy function whose minimization yields the optimal matching of superpixels images as well as their similarity. In our experiments we show that the proposed method is a useful tool for finding clusters in image sequences. Finally, we show that our approach outperforms state-of-the-art image comparison through non-rigid and affine matching of SURF features.
This is a preview of subscription content, log in via an institution to check access.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Preview
Unable to display preview. Download preview PDF.
Similar content being viewed by others
References
Ren, X., Malik, J.: Learning a classification model for segmentation. In: ICCV, pp. 10–17 (2003)
Gu, C., Lim, J., Arbelaez, P., Malik, J.: Recognition using regions. In: CVPR, pp. 1030–1037 (2009)
Vazquez-Reina, A., Avidan, S., Pfister, H., Miller, E.: Multiple hypothesis video segmentation from superpixel flows. In: Daniilidis, K., Maragos, P., Paragios, N. (eds.) ECCV 2010, Part V. LNCS, vol. 6315, pp. 268–281. Springer, Heidelberg (2010)
Wang, S., Lu, H., Yang, F., Yang, M.H.: Superpixel tracking. In: ICCV (2011)
Boltz, S., Nielsen, F., Soatto, S.: Earth mover distance on superpixels. In: ICIP, pp. 4597–4600 (2010)
Levinshtein, A., Stere, A., Kutulakos, K., Fleet, D., Dickinson, S., Siddiqi, K.: Turbopixels: Fast superpixels using geometric flows. IEEE Trans. Pattern Anal. Mach. Intell. 31(12), 2290–2297 (2009)
Pennec, X., Fillard, P., Ayache, N.: A riemannian framework for tensor computing. International Journal of Computer Vision 66(1), 41–66 (2006)
Zhang, F., Hancock, E.: New riemannian techniques for directional and tensorial image data. Pattern Recognition 43(4), 1590–1606 (2010)
Myronenko, A., Song, X.B.: Point-set registration: Coherent point drift. EEE Trans. on Pattern Analysis and Machine Intelligence 32(12), 2262–2275 (2010)
Tenenbaum, J.B., de Silva, V., Langford, J.C.: A global geometric framework for nonlinear dimensionality reduction. Science 290(5500), 2319–2323 (2000)
Pennec, X., Stefanescu, R., Arsigny, V., Fillard, P., Ayache, N.: Riemannian elasticity: A statistical regularization framework for non-linear registration. In: MICCAI, vol. 2, pp. 943–950 (2005)
Chiang, M.C., Leow, A., Klunder, A., Dutton, R., Barysheva, M., Rose, S., McMahon, K., de Zubicaray, G., Toga, A., Thompson, P.: Fluid registration of diffusion tensor images using information theory. IEEE Trans. Med. Imaging 27(4), 442–456 (2008)
Henze, N., Penrose, M.: On the multi-variate runs test. Annals of Statistics 27, 290–298 (1999)
Friedman, J., Rafsky, L.: Mutivariate generalization of the wald-wolfowitz and smirnov two-sample tests. Annals of Statistics 7(4), 697–717 (1979)
Stowell, D., Plumbley, M.: Fast multidimensional entropy estimation by k-d partitioning. IEEE Signal Processing Letters 16(6), 537–540 (2009)
Escolano, F., Hancock, E., Lozano, M.: Graph matching through entropic manifold alignment. In: CVPR, pp. 2417–2424 (2011)
Leonenko, N., Pronzato, L., Savani, V.: A class of renyi information estimators for multidimensional densities. Annals of Statistics 36(5), 2153–2182 (2008)
Bay, H., Ess, A., Tuytelaars, T., Gool, L.J.V.: Speeded-up robust features (surf). Computer Vision and Image Understanding 110(3), 346–359 (2008)
Author information
Authors and Affiliations
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2013 Springer-Verlag Berlin Heidelberg
About this paper
Cite this paper
Escolano, F., Hancock, E.R., Bonev, B., Lozano, M.A. (2013). TurboTensors for Entropic Image Comparison. In: Kropatsch, W.G., Artner, N.M., Haxhimusa, Y., Jiang, X. (eds) Graph-Based Representations in Pattern Recognition. GbRPR 2013. Lecture Notes in Computer Science, vol 7877. Springer, Berlin, Heidelberg. https://doi.org/10.1007/978-3-642-38221-5_6
Download citation
DOI: https://doi.org/10.1007/978-3-642-38221-5_6
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-642-38220-8
Online ISBN: 978-3-642-38221-5
eBook Packages: Computer ScienceComputer Science (R0)