K. Atsuko and S. Komatsu: Reconstruction of Blurred Image with Shift Variant PSF Using the SIngular Value Decomposition, Extended Abstracts (Optics Japan 1996); The Optical Society of Japan, The Japan Society of Applied Physics, pp. 35-36.

Abstract: We propose a new blind deconvolution algorithm using matrices which is available for thae case in which the PSF is shift variant. The conventional iterative Fourier transform blind deconvolution algorithm is effective when the PSF is shift invariant. The proposed new algorithm is an iterative method based on the pseudoinverse matrix calculation using the singular value decomposition. Image reconstruction from several blurred images were performed with this algorithm.