Image deconvolution using a stochastic differential equation approach

We consider the problem of image deconvolution. We foccus on a Bayesian approach which consists of maximizing an energy obtained by a Markov Random Field modeling. MRFs are classically optimized by a MCMC sampler embedded into a simulated annealing scheme. In a previous work, we have shown that, in the context of image denoising, a diffusion process can outperform the MCMC approach in term of computational time. Herein, we extend this approach to the case of deconvolution. We will first study the case where the kernel is known. Then, we will address the blind deconvolution.

• Type: Conference
• Year: 2007
• Where: VISAPP
• Authors: Xavier Descombes, M. Lebellego, Elena Zhizhina
• Keywords: Image Deconvolution, Markov Random Field, Simulated Annealing Scheme, VISAPP 2007, VISAPP 2008