A New Design Technology for Digital Image Magnifying Based on Hidden Markov Model
暂无分享,去创建一个
In this paper we propose a new design technology for digital image magnifying based on Hidden Markov Model (HMM). First, we focus on the corresponding rules between state sequence and the matrix of observations, and establish the topological model for digital images in the basis of pseudo-two-dimensional structure. By introducing the estimation of related possibilities, we propose a parameter learning algorithm combined with data smoothing method, which is specially designed for classified digital image processing. To verify the algorithm’s applicability and its actual result of image magnifying, we use Viterbi algorithm to Implement our newly proposed algorithm, and evaluate its performance in comparison with traditional interpolation methods. Compared to previous research, we expand the application category of HMM and provide new approaches for further research on image magnifying.
[1] Mark J. F. Gales,et al. The Application of Hidden Markov Models in Speech Recognition , 2007, Found. Trends Signal Process..
[2] P. Laplace. A Philosophical Essay On Probabilities , 1902 .