文件列表:

FTVd_v4.0 (0, 2009-07-05)
FTVd_v4.0\demoAll.m (201, 2009-07-19)
FTVd_v4.0\demoMTVL1.m (2882, 2009-07-05)
FTVd_v4.0\demoMTVL2.m (2565, 2009-07-05)
FTVd_v4.0\demoTVL1.m (1456, 2009-07-05)
FTVd_v4.0\demoTVL2.m (1313, 2009-07-05)
FTVd_v4.0\images (0, 2009-07-05)
FTVd_v4.0\images\lena256.png (122250, 2008-06-02)
FTVd_v4.0\READ_FTVd_v3.0.txt (12298, 2008-08-02)
FTVd_v4.0\solvers (0, 2009-07-05)
FTVd_v4.0\solvers\coresolvers (0, 2009-07-05)
FTVd_v4.0\solvers\coresolvers\ADM2MTVL1.m (9224, 2009-07-05)
FTVd_v4.0\solvers\coresolvers\ADM2MTVL2.m (8793, 2009-07-05)
FTVd_v4.0\solvers\coresolvers\ADM2TVL1.m (5786, 2009-07-05)
FTVd_v4.0\solvers\coresolvers\ADM2TVL2.m (5263, 2009-07-05)
FTVd_v4.0\solvers\FTVd_v4.m (1175, 2009-07-05)
FTVd_v4.0\solvers\utilities (0, 2009-07-04)
FTVd_v4.0\solvers\utilities\imfilter33.m (2193, 2008-08-01)
FTVd_v4.0\solvers\utilities\snr.m (484, 2009-05-17)

************************************************************************ FTVd_v4: Fast Total Variation Deconvolution (version 4) ************************************************************************ Copyright (C) 2009 Junfeng Yang, Yin Zhang, Wotao Yin and Yilun Wang 1). Get Started =================== Run any demo code listed below: demoTVL2 : Demo TV/L2 solve demoMTVL2 : Demo multichannel TV/L2 solve demoTVL1 : Demo TV/L1 solve demoMTVL1 : Demo multichannel TV/L1 solve or run demoAll to demostrate all. 2). Introduction ==================== FTVd_v4 keeps all the features of FTVd_v3.0 and is a major upgrade. In this version, we implemented the Alternating Direction Method (ADM), which dates back to Glowinski and Marrocco (1975) and Gabay and Mercier (1976). Compared with ver 3.0, this version avoids the ill-conditioning caussed by large penalty paramter by considering the augmented Lagrangian and ADM. As a result, FTVd_v4 has a much simply parameter selection rule and behaves very efficiently and roubustly. For more introductory material about FTVd, see READ_FTVd_v3.0, which is a copy of README in FTVd_v3.0. 3). Usage ==================== FTVd_v4 is called in the following way: out = FTVd_v4(Bn,H,mu,str), where Bn -- a blurry and noiy observation, H -- a convolution kernel, mu -- Lagrangian multiplier (positive number) str -- a string used to specify a model. * str = 'L2' or 'l2' to specify a L2 model (for white noise), * str = 'L1' or 'l1' to specify a L1 model (for impulsive noise). Users can also specify algorithm paramters and call FTVd_v4 like out = FTVd_v4(Bn,H,mu,str,opts), where opts is a structure with some fields. For example, when L1 model is to be solved, users may specify fields listed below: % opts -- a structure containing algorithm parameters {default} * opst.beta1 : a positive constant {5} * opst.beta2 : a positive constant {20} * opst.gamma : a constant in (0,1.618] {1.618} * opst.maxitr : maximum iteration number {500} * opst.relchg : a small positive parameter which controls stopping rule of the code. When the relative change of X is less than opts.relchg, then the code stops. {1.e-3} * opts.print : print inter results or not {0} For L2 models, .beta1 and .beta2 should be replaced by one field .beta. For more explanation about FTVd, see each core solver in folder "coresolver" and also check READ_FTVd_v3.0. 4). Contact Information ======================= FTVd is available at: http://www.caam.rice.edu/~optimization/L1/ftvd/ Please feel free to e-mail the following authors with any comments or suggestions: Junfeng Yang, Depart. Math., Nanjing Univ., 5). Copyright Notice ==================== FTVd is free software; you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation; either version 3 of the License, or (at your option) any later version. This program is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details at .

近期下载者：

相关文件：

评论：[我要评论] [举报此文件]

收藏者：