论文部分内容阅读
A novel algorithm for Bayesian document segmentation is proposed based on the wavelet domain hidden Markov tree (HMT) model. Once the parameters of model are known, according to the sequential maximum a posterior probability (SMAP) rule, firstly, the like