中文说明：Blind image deconvolution is an ill-posed problem that requires regularization to solve. However, many common forms of image prior used in this setting have a major draw- back in that the minimum of the resulting cost function does not correspond to the true sharp solution. Accordingly, a range of additional methods are needed to yield good re- sults (Bayesian methods, adaptive cost functions, alpha- matte extraction and edge localization). In this paper we introduce a new type of image regularization which gives lowest cost for the true sharp image. This allows a very simple cost formulation to be used for the blind deconvolu- tion model, obviating the need for additional methods. Due to its simplicity the algorithm is fast and very robust. We demonstrate our method on real images with both spatially invariant and spatially varying blur.
English Description：
Blind image deconvolution is an ill-posed problem thatrequires regularization to solve. However, many commonforms of image prior used in this setting have a major draw-back in that the minimum of the resulting cost function doesnot correspond to the true sharp solution. Accordingly, arange of additional methods are needed to yield good re-sults (Bayesian methods, adaptive cost functions, alpha-matte extraction and edge localization). In this paper weintroduce a new type of image regularization which giveslowest cost for the true sharp image. This allows a verysimple cost formulation to be used for the blind deconvolu-tion model, obviating the need for additional methods. Dueto its simplicity the algorithm is fast and very robust. Wedemonstrate our method on real images with both spatiallyinvariant and spatially varying blur.