A bound optimization approach to wavelet-based image deconvolution

We address the problem of image deconvolution under I_p norm (and other) penalties expressed in the wavelet domain. We propose an algorithm based on the bound optimization approach; this approach allows deriving EM-type algorithms without using the concept of missing/hidden data. The algorithm has provable monotonicity both with orthogonal or redundant wavelet transforms. We also derive bounds on the l_p norm penalties to obtain closed form update equations for any p ∈ [0, 2]. Experimental results show that the proposed method achieves state-of-the-art performance.