TY - GEN
T1 - Hidden Markov tree modeling of complex wavelet transforms
AU - Choi, Hyeokho
AU - Romberg, Justin
AU - Baraniuk, Richard
AU - Kingsbury, Nick
N1 - Funding Information:
HC, JR, RB supported by NSF grants MIP-9457438 and CCR- 9973188. ONR grant N00014-99-1-0813, DARPNAFOSR, grant F49620-97-1-0513, and Texas Instruments. NK supported by the Rice-Trinity Exchange Program and the UK Engineering and Physical Sciences Research Council (EPSRC).
Funding Information:
The energy compaction implied by the 2Populations property has inspired many simple schemes for removing additive white Gaussian noise from signals and images. Since most signal energy will concentrate in a few (large) wavelet coefficients while HC, JR, RB supported by NSF grants MIP-9457438 and CCR-9973188. ONR grant N00014-99-1-0813, DARPNAFOSR, grant F49620-97-1-05 13, and Texas Instruments. Email: {choi, jrom, richb} @ece.rice.edu; Internet: www.dsp.rice.edu NK supported by the Rice-Trinity Exchange Program and the UK Engineering and Physical Sciences Research Council (EPSRC). Email: [email protected]:I nternet: www,eng.cam.ac.uk/-ngk
Publisher Copyright:
© 2000 IEEE.
PY - 2000
Y1 - 2000
N2 - Multiresolution signal and image models such as the hidden Markov tree aim to capture the statistical structure of smooth and singular (edgy) regions. Unfortunately, models based on the orthogonal wavelet transform suffer from shift-variance, making them less accurate and realistic. We extend the HMT modeling framework to the complex wavelet transform, which features near shift-invariance and improved angular resolution compared to the standard wavelet transform. The model is computationally efficient (with linear-Time computation and processing algorithms) and applicable to general Bayesian inference problems as a prior density for the data. In a simple estimation experiment, the complex wavelet HMT model outperforms a number of high-performance denoising algorithms, including redundant wavelet thresholding (cycle spinning) and the redundant HMT.
AB - Multiresolution signal and image models such as the hidden Markov tree aim to capture the statistical structure of smooth and singular (edgy) regions. Unfortunately, models based on the orthogonal wavelet transform suffer from shift-variance, making them less accurate and realistic. We extend the HMT modeling framework to the complex wavelet transform, which features near shift-invariance and improved angular resolution compared to the standard wavelet transform. The model is computationally efficient (with linear-Time computation and processing algorithms) and applicable to general Bayesian inference problems as a prior density for the data. In a simple estimation experiment, the complex wavelet HMT model outperforms a number of high-performance denoising algorithms, including redundant wavelet thresholding (cycle spinning) and the redundant HMT.
UR - http://www.scopus.com/inward/record.url?scp=0033693081&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=0033693081&partnerID=8YFLogxK
U2 - 10.1109/ICASSP.2000.861889
DO - 10.1109/ICASSP.2000.861889
M3 - Conference contribution
AN - SCOPUS:0033693081
T3 - ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings
SP - 133
EP - 136
BT - Signal Processing Theory and Methods I
PB - Institute of Electrical and Electronics Engineers Inc.
T2 - 25th IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2000
Y2 - 5 June 2000 through 9 June 2000
ER -