Equiangular Kernel Dictionary Learning with Applications to Dynamic Texture Analysis
Authors:
- Yuhui QuanYQYuhui Quan
- Chenglong BaoCBChenglong Bao
- Hui JiHJHui Ji
- 1 June 2016
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 308-316
Abstract
Most existing dictionary learning algorithms consider a linear sparse model, which often cannot effectively characterize the nonlinear properties present in many types of visual data, e.g. dynamic texture (DT). Such nonlinear properties can be exploited by the so-called kernel sparse coding. This paper proposed an equiangular kernel dictionary learning method with optimal mutual coherence to exploit the nonlinear sparsity of high-dimensional visual data. Two main issues are addressed in the proposed method: (1) coding stability for redundant dictionary of infinite-dimensional space, and (2) computational efficiency for computing kernel matrix of training samples of high-dimensional data. The proposed kernel sparse coding method is applied to dynamic texture analysis with both local DT pattern extraction and global DT pattern characterization. The experimental results showed its performance gain over existing methods.Keywords
This publication has 41 references indexed in Scilit:
- Incoherent dictionary learning for sparse representation based image denoising2014
- An adaptive dictionary learning approach for modeling dynamical textures2014
- Design of Non-Linear Kernel Dictionaries for Object RecognitionIEEE Transactions on Image Processing, 2013
- View-invariant dynamic texture recognition using a bag of dynamical systems2009
- Online Prediction of Time Series Data With KernelsIEEE Transactions on Signal Processing, 2008
- Dictionary Preconditioning for Greedy AlgorithmsIEEE Transactions on Signal Processing, 2008
- Dynamic texture recognition2005
- Greed is Good: Algorithmic Results for Sparse ApproximationIEEE Transactions on Information Theory, 2004
- Grassmannian frames with applications to coding and communicationApplied and Computational Harmonic Analysis, 2003
- Temporal texture modeling2002