Bregman_Matlab_demo
所属分类:压缩解压
开发工具:matlab
文件大小:42KB
下载次数:34
上传日期:2013-03-20 13:32:31
上 传 者:
Leo_Su
说明: bregman 压缩感知恢复算法,属于贪婪算法的一种
(bregman signal recovery algorithm)
文件列表:
Bregman_Matlab_demo (0, 2012-01-20)
Bregman_Matlab_demo\error_forgetting_and_cancellation_demo.m (6690, 2012-01-20)
Bregman_Matlab_demo\fpc_bb_for_Bregman (0, 2012-01-16)
Bregman_Matlab_demo\fpc_bb_for_Bregman\fpc_bb.m (10769, 2012-01-16)
Bregman_Matlab_demo\fpc_bb_for_Bregman\fpc_bb_opts.m (4969, 2007-03-19)
Bregman_Matlab_demo\fpc_bb_for_Bregman\fp_bb.m (10939, 2012-01-16)
Bregman_Matlab_demo\fp_for_Bregman (0, 2012-01-16)
Bregman_Matlab_demo\fp_for_Bregman\fp.m (6326, 2012-01-16)
Bregman_Matlab_demo\fp_for_Bregman\fp_opts.m (3116, 2009-07-28)
Bregman_Matlab_demo\gpsr_for_Bregman (0, 2012-01-16)
Bregman_Matlab_demo\gpsr_for_Bregman\GPSR_Basic.m (21922, 2008-01-15)
Bregman_Matlab_demo\gpsr_for_Bregman\GPSR_BB.m (23882, 2009-01-16)
Bregman_Matlab_demo\gpsr_for_Bregman\source url.txt (75, 2012-01-16)
Bregman_Matlab_demo\mosek_for_Bregman (0, 2012-01-18)
Bregman_Matlab_demo\mosek_for_Bregman\mosek_uncon_l1.m (2060, 2012-01-18)
Bregman_Matlab_demo\mosek_for_Bregman\test_run.m (433, 2012-01-18)
Bregman_Matlab_demo\myprint.m (628, 2012-01-18)
Bregman_Matlab_demo\sparsa_for_Bregman (0, 2012-01-16)
Bregman_Matlab_demo\sparsa_for_Bregman\soft.m (123, 2007-12-13)
Bregman_Matlab_demo\sparsa_for_Bregman\source url.txt (79, 2012-01-16)
Bregman_Matlab_demo\sparsa_for_Bregman\SpaRSA.m (27788, 2009-04-21)
----- Introduction -----
The m-file "error_forgetting_and_cancellation_demo.m" is a Matlab demo of
the error forgetting and error cancellation properties of Bregman iteration
applied to solving the basis pursuit problem
min ||x||_1, s.t. Ax = B.
Bregman iteration is based on solving a sequence of subproblems in the form
of
min mu*||x||_1 + (1/2) ||Ax - b^k||_2^2
where b^k is updated iteratively as b^{k+1} = b^k + (b - A x^k).
In the demo, each subproblem is solved by one of the several different
algorithms up to tolerance 1E-6. The sequence, however, converges to a
highly accurate solution of error with an error less than 1E-14.
----- Installation -----
This code requires Mosek, which is a commercial solver with free academic
lincense.
Please download Mosek from www.mosek.com, install Mosek with a license,
and add its Matlab toolbox to the Matlab search path.
You can test run Mosek by executing "test_run.m" under "./mosek_for_Bregman"
----- Contact -----
Wotao Yin (wotao.yin@rice.edu, Jan 16, 2012)
近期下载者:
相关文件:
收藏者: