Abstract:Sparse representation leads to an efficient way to approximately recover a signal by the linear composition of a few bases from a learnt dictionary based on which various successful applications have been achieved. However, in the scenario of data compression, its efficiency and popularity are hindered. It is because of the fact that encoding sparsely distributed coefficients may consume more bits for representing the index of nonzero coefficients. Therefore, introducing an accurate rate constraint in sparse coding and dictionary learning becomes meaningful, which has not been fully exploited in the context of sparse representation. According to the Shannon entropy inequality, the variance of Gaussian distributed data bound its entropy, indicating the actual bitrate can be well estimated by its variance. Hence, a globally variance-constrained sparse representation (GVCSR) model is proposed in this paper, where a variance-constrained rate term is introduced to the optimization process. Specifically, we employ the alternating direction method of multipliers (ADMMs) to solve the non-convex optimization problem for sparse coding and dictionary learning, both of them have shown the state-of-the-art rate-distortion performance for image representation. Furthermore, we investigate the potential of applying the GVCSR algorithm in the practical image set compression, where the optimized dictionary is trained to efficiently represent the images captured in similar scenarios by implicitly utilizing inter-image correlations. Experimental results have demonstrated superior rate-distortion performance against the state-of-the-art methods.

Efficient Sparsity Estimation Via Marginal-Lasso Coding.

Efficient Probabilistic Latent Semantic Analysis with Sparsity Control

Sparse Estimation Via ℓ_q Optimization Method in High-Dimensional Linear Regression

Local Hybrid Coding for Image Classification

Sparse Estimation Via Lower-Order Penalty Optimization Methods in High-Dimensional Linear Regression.

Efficient Network Compression Through Smooth-Lasso Constraint

Elastic Net with Adaptive Weight for Image Denoising

Diffusion Sparse Least-Mean Squares over Networks

Sparse Poisson coding for high dimensional document clustering

Sparsity and smoothness via the fused lasso

A Fast Method for Lasso and Logistic Lasso

Lasso with Latents: Efficient Estimation, Covariate Rescaling, and Computational-Statistical Gaps

Globally Variance-Constrained Sparse Representation and Its Application in Image Set Coding.

Efficient Marginalization of Discrete and Structured Latent Variables via Sparsity

Sparsity-Constraint Optimization via Splicing Iteration

Low-Rank Sparse Coding for Image Classification

Transfer Sparse Coding for Robust Image Representation

Topic Level Sampling Towards Optimized Locality Sensitive Vocabulary Coding

Fused LASSO Penalized Least Absolute Deviation Estimator for High Dimensional Linear Regression

Parameterizing Region Covariance: An Efficient Way To Apply Sparse Codes On Second Order Statistics

Sparsity aware coding for single photon sensitive vision using Selective Sensing