Blind Image Deconvolution Using Bussgang Techniques: Applications to Image Deblurring and Texture Synthesis
Patrizio Campisi, Alessandro Neri, Stefania Colonnese, Gianpiero Panci, Gaetano Scarano · 2017
In this chapter, Bussgang blind deconvolution techniques are reviewed in the general Bayesian framework of minimum mean square error (MMSE) estimation, and some recent activities of the authors on both single-channel and multichannel blind image deconvolution, under the general framework of Bussgang deconvolution, are described. Applications of the single-channel Bussgang blind deconvolution approach to perform unsupervised texture synthesis is detailed. Moreover, the Bussgang blind deconvolution method is generalized to the multichannel case with application to image deblurring problems. Specifically, we address the restoration problem of poorly spatially correlated images as well as strongly correlated (natural) images and experimental results pertaining to restoration of motion blurred text images, out-of-focus spiky images, and blurred natural images are given. A theoretical analysis to show the local convergence properties of the Bussgang deconvolution algorithm is also reported.