Learning Hierarchical Dictionary for Shape Patterns

Shape information is essential for image understanding. Decomposing images into shape patterns using a learned dictionary can provide an effective image representation. However, most of the dictionary based methods retain no structure information between dictionary elements. In this study, We propose Hierarchical Dictionary Shape Decomposition (HiDiShape) to learn a hierarchical dictionary for image shape patterns. Shift Invariant Sparse Coding and HMAX model are combined to decompose image into common shape patterns. And the Sparse Spatial and Hierarchical Regularization (SSHR) is proposed to organize these shape patterns to construct tree structured dictionary. Experiments show that the proposed HiDiShape method can learn tree structured dictionaries for complex shape patterns, and the hierarchical dictionaries improve the performances of corrupted shape reconstruction task.

[1]  David J. Field,et al.  Sparse coding with an overcomplete basis set: A strategy employed by V1? , 1997, Vision Research.

[2]  P. Zhao,et al.  The composite absolute penalties family for grouped and hierarchical variable selection , 2009, 0909.0411.

[3]  Michael Elad,et al.  Learning Multiscale Sparse Representations for Image and Video Restoration , 2007, Multiscale Model. Simul..

[4]  T. Poggio,et al.  Hierarchical models of object recognition in cortex , 1999, Nature Neuroscience.

[5]  Eric P. Xing,et al.  Tree-Guided Group Lasso for Multi-Task Regression with Structured Sparsity , 2009, ICML.

[6]  Rajat Raina,et al.  Self-taught learning: transfer learning from unlabeled data , 2007, ICML '07.

[7]  Marc'Aurelio Ranzato,et al.  Learning invariant features through topographic filter maps , 2009, CVPR.

[8]  Jonathan Le Roux,et al.  Adaptive Template Matching with Shift-Invariant Semi-NMF , 2008, NIPS.

[9]  Bo Zhang,et al.  SISCHMAX: Discovering common contour patterns , 2010, 9th IEEE International Conference on Cognitive Informatics (ICCI'10).

[10]  S.M. Plis,et al.  Sparse shift-invariant NMF , 2008, 2008 IEEE Southwest Symposium on Image Analysis and Interpretation.

[11]  Rajat Raina,et al.  Efficient sparse coding algorithms , 2006, NIPS.

[12]  J. P. Jones,et al.  An evaluation of the two-dimensional Gabor filter model of simple receptive fields in cat striate cortex. , 1987, Journal of neurophysiology.

[13]  Thomas Serre,et al.  Robust Object Recognition with Cortex-Like Mechanisms , 2007, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[14]  Andrew McCallum,et al.  Piecewise pseudolikelihood for efficient training of conditional random fields , 2007, ICML '07.

[15]  D. Hubel,et al.  Receptive fields, binocular interaction and functional architecture in the cat's visual cortex , 1962, The Journal of physiology.

[16]  Julien Mairal,et al.  Proximal Methods for Sparse Hierarchical Dictionary Learning , 2010, ICML.

[17]  David G. Lowe,et al.  University of British Columbia. , 1945, Canadian Medical Association journal.

[18]  Samy Bengio,et al.  Group Sparse Coding , 2009, NIPS.

[19]  Thomas Martinetz,et al.  Simple Method for High-Performance Digit Recognition Based on Sparse Coding , 2008, IEEE Transactions on Neural Networks.

[20]  Guillermo Sapiro,et al.  Online dictionary learning for sparse coding , 2009, ICML '09.

[21]  Dennis Gabor,et al.  Theory of communication , 1946 .