Image Denoising Based on Contourlet-Domain HMT Models Using Cycle Spinning

We propose a new method for image denoising based on contourlet-domain hidden Markov tree (CHMT) models, which have been recently introduced. CHMT models achieve superior denoising results over wavelet-domain HMT (WHMT) models in terms of visual quality. But denoising by means of CHMT still introduces some artifacts due to the lack of translation invariance of the contourlet transform. We employ a cycle-spinning-based technique to develop translation invariant CHMT denoising scheme. This scheme achieves enhanced estimation results for images that are corrupted with additive Gaussian noise. Our experiments show that the proposed approach outperforms both WHMT-based denoising method and CHMT-based denoising method, in both visual quality and the PSNR values.