Computational methods for hidden Markov tree models-an application to wavelet trees

The hidden Markov tree models were introduced by Crouse et al. in 1998 for modeling nonindependent, non-Gaussian wavelet transform coefficients. In their paper, they developed the equivalent of the forward-backward algorithm for hidden Markov tree models and called it the "upward-downward algorithm". This algorithm is subject to the same numerical limitations as the forward-backward algorithm for hidden Markov chains (HMCs). In this paper, adapting the ideas of Devijver from 1985, we propose a new "upward-downward" algorithm, which is a true smoothing algorithm and is immune to numerical underflow. Furthermore, we propose a Viterbi-like algorithm for global restoration of the hidden state tree. The contribution of those algorithms as diagnosis tools is illustrated through the modeling of statistical dependencies between wavelet coefficients with a special emphasis on local regularity changes.

[1]  Richard G. Baraniuk,et al.  Multiscale image segmentation using wavelet-domain hidden Markov models , 2001, IEEE Trans. Image Process..

[2]  G. Churchill.  Stochastic models for heterogeneous DNA sequences. , 1989, Bulletin of mathematical biology.

[3]  L. Baum,et al.  A Maximization Technique Occurring in the Statistical Analysis of Probabilistic Functions of Markov Chains , 1970 .

[4]  Paolo Frasconi,et al.  Image Document Categorization Using Hidden Tree Markov Models and Structured Representations , 2001, ICAPR.

[5]  L. R. Rabiner,et al.  An introduction to the application of the theory of probabilistic functions of a Markov process to automatic speech recognition , 1983, The Bell System Technical Journal.

[6]  Frederick Jelinek,et al.  Statistical methods for speech recognition , 1997 .

[7]  Lawrence Carin,et al.  Dual hidden Markov model for characterizing wavelet coefficients from multi-aspect scattering data , 2001, Signal Process..

[8]  John B. Moore,et al.  A Soft Output Hybrid Algorithm for ML/MAP Sequence Estimation , 1998, IEEE Trans. Inf. Theory.

[9]  S. Mallat.  A wavelet tour of signal processing , 1998 .

[10]  R. Peltier,et al.  Multifractional Brownian Motion : Definition and Preliminary Results , 1995 .

[11]  G. McLachlan,et al.  The EM algorithm and extensions , 1996 .

[12]  Gregory W. Wornell,et al.  Estimation of fractal signals from noisy measurements using wavelets , 1992, IEEE Trans. Signal Process..

[13]  S. Jaffard.  Pointwise smoothness, two-microlocalization and wavelet coefficients , 1991 .

[14]  Neri Merhav,et al.  Hidden Markov processes , 2002, IEEE Trans. Inf. Theory.

[15]  Lawrence R. Rabiner,et al.  A tutorial on hidden Markov models and selected applications in speech recognition , 1989, Proc. IEEE.

[16]  Michael I. Jordan,et al.  Probabilistic Independence Networks for Hidden Markov Probability Models , 1997, Neural Computation.

[17]  Robert D. Nowak,et al.  Wavelet-based statistical signal processing using hidden Markov models , 1998, IEEE Trans. Signal Process..

[18]  Patrick Flandrin,et al.  Wavelet analysis and synthesis of fractional Brownian motion , 1992, IEEE Trans. Inf. Theory.

[19]  Pierre A. Devijver,et al.  Baum's forward-backward algorithm revisited , 1985, Pattern Recognit. Lett..

[20]  J. R. Rohlicek,et al.  Parameter estimation of dependence tree models using the EM algorithm , 1995, IEEE Signal Processing Letters.

[21]  Patrice Abry,et al.  Wavelet Analysis of Long-Range-Dependent Traffic , 1998, IEEE Trans. Inf. Theory.