This project is for performing compressed sensing reconstruction of sparse signal via the BHT-BP solver.
README for BHT_BP solver by Jaewook Kang (jwkkang@gist.ac.kr)
This package contains methods for performing compressed sensing reconstruction of sparse signal via the BHT-BP solver.
-Jaewook Kang, Heung-No Lee, Kiseon Kim, "Bayesian Hypothesis Test using Nonparametric Belief Propagation for Noisy Sparse Recovery," IEEE Trans. on Signal process., vol. 63, no. 4, pp. 935-948, Feb. 2015
This package shows a numerical comparison of solvers for compressed sensing recovery. The solvers included in this comparison are as given below:
-CS-BP : D. Baron, S. Sarvotham, and R. Baraniuk, "Bayesian compressive sensing via belief propagation,
IEEE Trans. Signal Process., vol. 58, no. 1, pp. 269-280, Jan. 2010.
-BCS: Shihao Ji, Ya Xue, and Lawrence Carin, "Bayesian compressive sensing,"
IEEE Trans. Signal Process., vol. 56, no. 6, pp. 2346-2356, June. 2008.
-SuPrEM: M. Akcakaya, J. Park, and V. Tarokh, “A coding theory approach to noisy compressive
sensing using low density frame,” IEEE Trans. Signal Process., vol. 59, no. 12, pp. 5369-5379, Nov. 2011.
-L1-DS: E. Candes and T. Tao, “The Dantzig selector: Statistical estimation when p is much larger than n,”
Ann. Statist., vol. 35, no. 6, pp. 2313?2351, 2007
==========================================================================
The BHT-BP solver includes many input and output parameter, which is described below. For the other solvers, please see each author's website or the corresponding papers.
[estX_BHT,estX_CSBP,Merror_CSBP,Merror_BHT,state_err_BHT,state_err_CSBP,telapsed] =BHT_BP_ver3(trueX,H,Z,q,maxiter,sigma_N,sigmaX,Xstate,Nd,Xmin)
- trueX - This is the true value of the sparse signal X, which is used to calculate MSE performance of the reconstruction
- H - This is a measurement matrix (M by N)
- Z- This is a noisy measurment vector (M by 1).
- q - The sparsity rate which is a parameter of the prior PDF
- maxiter - The maxmum number of belief propagation iterations in BHT-BP
- sigma_N - The standard deviation for additive measurement noise
- sigma_X - The standard deviation for signal, which is a parameter of the prior PDF
- Xstate - The binary state vector indicating the location of nonzero value in the signal.
- estX_BHT- The BHT-BP signal estimate
- estX_CSBP -The CS-BP signal estimate
- Merror_CSBP - The MSE performance of the BHT-BP recovery
- Merror_BHT - The MSE performance of the CS-BP recovery
- state_err_BHT- The number of state errors in the detected signal support by the BHT-BP recovery
- state_err_CSBP - The number of state errors in the detected signal support by the CS-BP recovery
- telapsed - the running time expending for the belief propagation iterations =========================================================================
Size: The total size of the file is 1.12 MB Player Information: " MATLAB Version: 8.2.0.701 (R2013b) " Operating System: Microsoft Windows 7 Version 6.1 (Build 7601: Service Pack 1) " Java Version: Java 1.7.0_11-b21 with Oracle Corporation Java HotSpot(TM) 64-Bit Server VM mixed mode
Packing List:
demo_BHT_BP.m: provides an exemplary sparse signal recovery among BHT-BP, CS-BP, and BCS solvers.
demo_ImagRev.m: providing image (cameraman128x128) recovery demonstration using BHT-BP, CS-BP, and BCS solvers.
BHT_BP_Fig10_MSE_wrt_Nd_ver1.m: The experment setup for Fig 10 in the corresponding paper.
BHT_BP_Fig11a_MSE_wrt_L_M512.m : The experment setup for Fig 11-(a) in the corresponding paper.
BHT_BP_Fig11b_MSE_wrt_L_M768.m : The experment setup for Fig 11-(a) in the corresponding paper.
BHT_BP_Fig8n9_MSE_wrt_SNR_ver2.m: The experment setup for Fig 8 and Fig 9 in the corresponding paper.
Contact Information:
" Lab. phone: +82-62-715-2264 " E-mail : jwkkang@gist.ac.kr, jwkang10@gmail.com " Final Update @ May 2015