Learn how easy it is to sync an existing GitHub or Google Code repo to a SourceForge project!

## RE: [Vxl-users] Problems with estimating errors in vnl_levenberg_marquardt

 RE: [Vxl-users] Problems with estimating errors in vnl_levenberg_marquardt From: Peter Vanroose - 2003-08-28 09:18:40 ```> You'll have to go into the LMDIF routine upon which > vnl_levenberg_marquardt is built > and recover the Jacobian from in there. Would it be worthwile to add an interface to the Jacobian in the C++ class vnl_levenberg_marquardt ? -- Peter. ```

 [Vxl-users] Problems with estimating errors in vnl_levenberg_marquardt From: Stanislaw Szpala - 2003-08-27 20:16:03 Attachments: Message as HTML ```Hi, I'm using vnl_levenberg_marquardt class to fit my data set (I supply no = gradients). I run the code from Microsoft Visual C++ 6.0; no MFC = classes, Windows2000. It works, but I need to get an estimate of the = errors of the fitted parameters. I believe the errors can be taken from = the diagonal elements of the covariance matrix. Unfortunately, I am = unable to get the covariance matrix using get_covariance(), see below.=20 ... vnl_vector x0(3); ... vnl_vector x1 =3D x0; vnl_levenberg_marquardt lm(fittedFunction); lm.minimize_without_gradient(x1); vnl_matrix covar(3,3); covar =3D lm.get_covariance(); ('fitedFunction' is my class derived from 'vnl_least_squares_function', = and it defines the fitted function 'f') When I try accesing get_JtJ(), I get "not implemented yet" message, see: jacobian =3D lm.get_JtJ(); So it looks like it is not implemented. But the Jacobians must be = calculated at some point. So, how can I access the covariance matrix? Thanks, Stan Stanislaw Szpala, Ph.D. Robarts Research Institute 100 Perth Drive London Ontario N6A 5K8, Canada tel (519)663-5777x34213 sszpala@...```
 RE: [Vxl-users] Problems with estimating errors in vnl_levenberg_marquardt From: Andrew Fitzgibbon - 2003-08-28 09:08:25 ```Hi, You'll have to go into the LMDIF routine upon which vnl_levenberg_marquardt is built and recover the Jacobian from in there. If you know FORTRAN this should be fairly easy, and if you don't, it might take a little time. Do we have any FORTRAN speakers on the list? A. -----Original Message----- From: vxl-users-admin@... [mailto:vxl-users-admin@...] On Behalf Of Stanislaw Szpala Sent: 27 August 2003 21:18 To: vxl-users@... Subject: [Vxl-users] Problems with estimating errors in vnl_levenberg_marquardt Hi, I'm using vnl_levenberg_marquardt class to fit my data set (I supply no gradients). I run the code from Microsoft Visual C++ 6.0; no MFC classes, Windows2000. It works, but I need to get an estimate of the errors of the fitted parameters. I believe the errors can be taken from the diagonal elements of the covariance matrix. Unfortunately, I am unable to get the covariance matrix using get_covariance(), see below. ... vnl_vector x0(3); ... vnl_vector x1 = x0; vnl_levenberg_marquardt lm(fittedFunction); lm.minimize_without_gradient(x1); vnl_matrix covar(3,3); covar = lm.get_covariance(); ('fitedFunction' is my class derived from 'vnl_least_squares_function', and it defines the fitted function 'f') When I try accesing get_JtJ(), I get "not implemented yet" message, see: jacobian = lm.get_JtJ(); So it looks like it is not implemented. But the Jacobians must be calculated at some point. So, how can I access the covariance matrix? Thanks, Stan Stanislaw Szpala, Ph.D. Robarts Research Institute 100 Perth Drive London Ontario N6A 5K8, Canada tel (519)663-5777x34213 sszpala@... ```
 RE: [Vxl-users] Problems with estimating errors in vnl_levenberg_marquardt From: Peter Vanroose - 2003-08-28 09:18:40 ```> You'll have to go into the LMDIF routine upon which > vnl_levenberg_marquardt is built > and recover the Jacobian from in there. Would it be worthwile to add an interface to the Jacobian in the C++ class vnl_levenberg_marquardt ? -- Peter. ```
 RE: [Vxl-users] Problems with estimating errors in vnl_levenberg_marquardt From: Andrew Fitzgibbon - 2003-08-28 10:06:05 ```That's a very good plan -- there's a stub ther which Stanislaw mentioned, but it's not connected to anything. > -----Original Message----- > From: vxl-users-admin@... > [mailto:vxl-users-admin@...] On Behalf Of > Peter Vanroose > Sent: 28 August 2003 10:18 > To: Andrew Fitzgibbon > Cc: 'Stanislaw Szpala'; 'VXL Users List (E-mail)' > Subject: RE: [Vxl-users] Problems with estimating errors in > vnl_levenberg_marquardt > > > > You'll have to go into the LMDIF routine upon which > > vnl_levenberg_marquardt is built > > and recover the Jacobian from in there. > > Would it be worthwile to add an interface to the Jacobian > in the C++ class vnl_levenberg_marquardt ? > > > -- Peter. > > > ------------------------------------------------------- > This sf.net email is sponsored by:ThinkGeek > Welcome to geek heaven. > http://thinkgeek.com/sf > _______________________________________________ > Vxl-users mailing list > Vxl-users@... > https://lists.sourceforge.net/lists/listinfo/vxl-users > ```