Variational Bayesian image restoration with multi-structured model of wavelet transform coefficients

作者：

Highlights：

• A multi-structured wavelet Bayesian compressive sensing (CS) model is proposed.

• This model is used for image compression under the discrete wavelet transform.

• The multi-structured prior, including the cluster and tree structures, is used.

• Variational Bayesian inference is used to efficiently derive the model parameters.

• The experiments show that our proposed model outperforms than other CS algorithms.

摘要

•A multi-structured wavelet Bayesian compressive sensing (CS) model is proposed.•This model is used for image compression under the discrete wavelet transform.•The multi-structured prior, including the cluster and tree structures, is used.•Variational Bayesian inference is used to efficiently derive the model parameters.•The experiments show that our proposed model outperforms than other CS algorithms.

论文关键词：Bayesian compressive sensing,Wavelet-tree structure,Multi-structure model,Variational Bayes inference

论文评审过程：Received 26 April 2018, Revised 7 November 2018, Accepted 4 December 2018, Available online 6 December 2018, Version of Record 17 December 2018.

论文官网地址：https://doi.org/10.1016/j.image.2018.12.003