Image deconvolution using Hidden Markov Tree modeling of complex wavelet packets

A. Jalobeanu, N. Kingsbury and J. Zerubia

IEEE International Conference on Image Processing, ICIP2001, Greece.

Abstract: In this paper, we propose to use a hidden Markov tree mod­ eling of the complex wavelet packet transform, to capture the inter­scale dependencies of natural images. First, the observed image, blurred and noisy, is deconvolved with­ out regularization. Then its transform is denoised within a Bayesian framework using the proposed model, whose pa­ rameters are estimated by an EM technique. The total com­ plexity of this new deblurring algorithm remains O(N).