Improved Hidden Markov Tree Model and Its Application in Image Denoising
Wenzhe Lu, Zhe Zhang, Xixian Niu · 2008
Wavelet coefficients of natural images have significant dependencies. HMT model only capture the interscale dependencies. In this paper we propose an improved HMT model, which simply model intrascale dependencies as multivariate Gaussian distribution. This model is also applied to image denoising. Experimental results show that it can provide better performance than HMT model.