A novel framework method for non-blind deconvolution using subspace images priors

作者：

Highlights：

• A novel framework for non-blind deconvolution is proposed.

• The priors of image subspaces are utilized to improve performance.

• Precise image structures can be protected for performance promotion.

• The framework can be generally extended to existing methods.

• The framework can be more robust to noise by subspaces decomposition.

摘要

Highlights•A novel framework for non-blind deconvolution is proposed.•The priors of image subspaces are utilized to improve performance.•Precise image structures can be protected for performance promotion.•The framework can be generally extended to existing methods.•The framework can be more robust to noise by subspaces decomposition.

论文关键词：Non-blind deconvolution,Subspace images priors,Existing deblurring techniques,Least square integration

论文评审过程：Received 9 November 2015, Revised 14 April 2016, Accepted 14 April 2016, Available online 28 April 2016, Version of Record 11 May 2016.

论文官网地址：https://doi.org/10.1016/j.image.2016.04.003