|
Size: 1034
Comment:
|
Size: 3195
Comment:
|
| Deletions are marked like this. | Additions are marked like this. |
| Line 5: | Line 5: |
| *http://kutsyy.com/java/javadoc/index.html -- Java libraries for several needed functions = files modified/to-be-modified = == utils/eigen.c == * will be removed == utils/image.c == * EigenSystem will be replaced == utils/utils.c == * ran1 will be replaced == utils/mrinorm.c == * spline will be replaced * splint will be replaced == utils/matrix.c == * lu_matrix_inverse * MatrixDeterminant will be replaced with OpenMatrixDeterminant * svdcmp |
|
| Line 9: | Line 29: |
| *itkPowellOptimizer *vnl_powell == f1dim.c == *function used by linmin [10.5] |
|
| Line 12: | Line 37: |
| *nothing apparent in ITK, vnl, GSL *http://cvs.gnome.org/viewcvs/moiss/moiss-src/linmin.c *check koders.com *used by powell |
|
| Line 15: | Line 44: |
| *vnl_rnpoly_solve has method | |
| Line 18: | Line 48: |
| *vnl_rnpoly_solve has method | |
| Line 21: | Line 52: |
| *gsl_min_fminimizer_brent *ITK/vnl: Utilities/vxl/core/vnl/algo/vnl_brent.cxx *used by linmin |
|
| Line 22: | Line 56: |
| == mbrak.c == | == mnbrak.c == |
| Line 24: | Line 58: |
| *no apparent replacements. others have same conclusion: http://www.blackwellpublishing.com/rss/Readmefiles/zucker.htm *check koders.com *used by linmin |
|
| Line 27: | Line 64: |
| *LAPACK seems to have it (http://www.netlib.org/lapack/lug/node53.html), but not ATLAS *check koders.com |
|
| Line 29: | Line 68: |
| *random deviate, minimal standard plus shuffle | *random deviate, minimal standard plus shuffle [7.1] * '''used in''' *utils/utils.c from within randomNumber() *not supposed to be such a great random num generator. how important is using the same algorithm? *exact same: gsl_rng_ran1, but no longer exists. now provides gsl_rng_minstd *full gsl doc for this: http://www.lsw.uni-heidelberg.de/manuals/gsl-ref-html/gsl-ref_16.html *vnl_random *itkMersenneTwisterRandomVariateGenerator |
| Line 33: | Line 79: |
| *itkSymmetricEigenAnalysis has ReduceToTridiagonalMatrixAndGetTransformation, adapted from netlib/tred2.c *also: Utilities/vxl/v3p/netlib/tred2.c |
|
| Line 36: | Line 84: |
| *tqli vs. tql1 -- the same I think *itkSymmetricEigenAnalysis has ComputeEigenValuesUsingQL, adapted from netlib/tql1.c *also: Utilities/vxl/v3p/netlib/tql1.c |
|
| Line 39: | Line 90: |
| *Image Registration Toolkit appears to have closed source implementation. Maybe they'll share. *SPLINE library: http://www.scs.fsu.edu/~burkardt/cpp_src/spline/spline.html |
|
| Line 45: | Line 98: |
| *ITK: Utilities/vxl/v3p/netlib/pythag.c *used by tqli |
|
| Line 48: | Line 103: |
| *check koders.com | |
| Line 50: | Line 106: |
| *search along a line, used by newt | *search along a line, used by dfpmin *check koders.com |
| Line 53: | Line 110: |
| * | *vector, matrix, free_vector, free_matrix |
general links
http://www.numerical-recipes.com/recipe-list.html -- NR method names descriptions
http://kutsyy.com/java/javadoc/index.html -- Java libraries for several needed functions
files modified/to-be-modified
utils/eigen.c
- will be removed
utils/image.c
EigenSystem will be replaced
utils/utils.c
- ran1 will be replaced
utils/mrinorm.c
- spline will be replaced
- splint will be replaced
utils/matrix.c
- lu_matrix_inverse
MatrixDeterminant will be replaced with OpenMatrixDeterminant
- svdcmp
specific functions
powell.c
- powell's direction set method, or powell minimizer
- itkPowellOptimizer
- vnl_powell
f1dim.c
- function used by linmin [10.5]
linmin.c
- line minimization
- nothing apparent in ITK, vnl, GSL
- check koders.com
- used by powell
ludcmp.c
- LU decomposition
- vnl_rnpoly_solve has method
lubksb.c
- LU back substitution
- vnl_rnpoly_solve has method
brent.c
- brent's method for min of function
- gsl_min_fminimizer_brent
- ITK/vnl: Utilities/vxl/core/vnl/algo/vnl_brent.cxx
- used by linmin
mnbrak.c
- bracket the minimum of a function
no apparent replacements. others have same conclusion: http://www.blackwellpublishing.com/rss/Readmefiles/zucker.htm
- check koders.com
- used by linmin
svdcmp.c
- singular value decomposition of a matrix
LAPACK seems to have it (http://www.netlib.org/lapack/lug/node53.html), but not ATLAS
- check koders.com
ran1.c
- random deviate, minimal standard plus shuffle [7.1]
used in
- utils/utils.c from within randomNumber()
- not supposed to be such a great random num generator. how important is using the same algorithm?
- exact same: gsl_rng_ran1, but no longer exists. now provides gsl_rng_minstd
full gsl doc for this: http://www.lsw.uni-heidelberg.de/manuals/gsl-ref-html/gsl-ref_16.html
- vnl_random
- itkMersenneTwisterRandomVariateGenerator
tred2.c
- householder reduction of a real, symmetric matrix
itkSymmetricEigenAnalysis has ReduceToTridiagonalMatrixAndGetTransformation, adapted from netlib/tred2.c
- also: Utilities/vxl/v3p/netlib/tred2.c
tqli.c
- eigensolution of a symmetric tridiagonal matrix
- tqli vs. tql1 -- the same I think
- itkSymmetricEigenAnalysis has ComputeEigenValuesUsingQL, adapted from netlib/tql1.c
- also: Utilities/vxl/v3p/netlib/tql1.c
splint.c
- cubic spline interpolation [3.3]
- Image Registration Toolkit appears to have closed source implementation. Maybe they'll share.
SPLINE library: http://www.scs.fsu.edu/~burkardt/cpp_src/spline/spline.html
spline.c
- construct a cubic spline [3.3]
pythag.c
calculate (a2+b2)^{1/2} without overflow [2.6]
- ITK: Utilities/vxl/v3p/netlib/pythag.c
- used by tqli
dfpmin.c
- minimize in N-dimensions by variable metric method
- check koders.com
lnsrch.c
- search along a line, used by dfpmin
- check koders.com
nrutil.c
- vector, matrix, free_vector, free_matrix
