2013/03/15 by Se Un Park, Park, Se Un, Nicolas Dobigeon +3
Engineering · Medicine · #Advanced MRI Techniques and Applications #Data Analysis #FOS: Computer and information sciences #FOS: Physical sciences #Machine Learning (stat.ML) #Photoacoustic and Ultrasonic Imaging #Sparse and Compressive Sensing Techniques #Statistics and Probability (physics.data-an)
paper · pdf · doi:10.48550/arxiv.1303.3866
openalex publication_date 2013/03/15 · openalex created_date 2022/09/19 · openalex updated_date 2026/07/28
We present a variational Bayesian method of joint image reconstruction and\npoint spread function (PSF) estimation when the PSF of the imaging device is\nonly partially known. To solve this semi-blind deconvolution problem, prior\ndistributions are specified for the PSF and the 3D image. Joint image\nreconstruction and PSF estimation is then performed within a Bayesian\nframework, using a variational algorithm to estimate the posterior\ndistribution. The image prior distribution imposes an explicit atomic measure\nthat corresponds to image sparsity. Importantly, the proposed Bayesian\ndeconvolution algorithm does not require hand tuning. Simulation results\nclearly demonstrate that the semi-blind deconvolution algorithm compares\nfavorably with previous Markov chain Monte Carlo (MCMC) version of myopic\nsparse reconstruction. It significantly outperforms mismatched non-blind\nalgorithms that rely on the assumption of the perfect knowledge of the PSF. The\nalgorithm is illustrated on real data from magnetic resonance force microscopy\n(MRFM).\n