contrib/mul/mbl/mbl_matrix_products.h File Reference

Various specialised versions of matrix product operations. More...

#include <vnl/vnl_fwd.h>

Go to the source code of this file.

Functions
void	mbl_matrix_product (vnl_matrix< double > &AB, const vnl_matrix< double > &A, const vnl_matrix< double > &B)
	Compute product AB = A * B.
void	mbl_matrix_product_a_bt (vnl_matrix< double > &ABt, const vnl_matrix< double > &A, const vnl_matrix< double > &B)
	Compute product ABt = A * B.transpose().
void	mbl_matrix_product_a_bt (vnl_matrix< double > &ABt, const vnl_matrix< double > &A, const vnl_matrix< double > &B, int n_cols)
	Compute ABt = A * B.transpose() using only first n_cols columns of A and B.
void	mbl_matrix_product_a_at (vnl_matrix< double > &AAt, const vnl_matrix< double > &A, unsigned nr, unsigned nc)
	Compute AAt = A * A.transpose(), using only first nr x nc partition of A.
void	mbl_matrix_product_a_at (vnl_matrix< double > &AAt, const vnl_matrix< double > &A)
	Compute AAt = A * A.transpose().
void	mbl_matrix_product_at_b (vnl_matrix< double > &AtB, const vnl_matrix< double > &A, const vnl_matrix< double > &B)
	Compute product AtB = A.transpose() * B.
void	mbl_matrix_product_at_b (vnl_matrix< double > &AtB, const vnl_matrix< double > &A, const vnl_matrix< double > &B, int ncols_a)
	Compute AtB = A.transpose() * B , using first ncols_a cols of A.
void	mbl_matrix_product_at_a (vnl_matrix< double > &AtA, const vnl_matrix< double > &A, unsigned nc)
	Compute product AtA = A.transpose() * A using nc cols of A.
void	mbl_matrix_product_at_a (vnl_matrix< double > &AtA, const vnl_matrix< double > &A)
	Compute product AtA = A.transpose() * A.
void	mbl_matrix_product_adb (vnl_matrix< double > &ADB, const vnl_matrix< double > &A, const vnl_vector< double > &d, const vnl_matrix< double > &B)
	ADB = A * D * B where D is diagonal with elements d.

---

Detailed Description

Various specialised versions of matrix product operations.

Author:

Tim Cootes

Date:

25-Apr-2001

Definition in file mbl_matrix_products.h.

---

Function Documentation

void mbl_matrix_product	(	vnl_matrix< double > &	AB,
		const vnl_matrix< double > &	A,
		const vnl_matrix< double > &	B
	)

Compute product AB = A * B.

Definition at line 16 of file mbl_matrix_products.cxx.

void mbl_matrix_product_a_at	(	vnl_matrix< double > &	AAt,
		const vnl_matrix< double > &	A,
		unsigned	nr,
		unsigned	nc
	)

Compute AAt = A * A.transpose(), using only first nr x nc partition of A.

Uses symmetry of result to improve speed

Definition at line 122 of file mbl_matrix_products.cxx.

void mbl_matrix_product_a_at	(	vnl_matrix< double > &	AAt,
		const vnl_matrix< double > &	A
	)

Compute AAt = A * A.transpose().

Uses symmetry of result to improve speed

Compute AAt = A * A.transpose().

Uses symmetry of result to be approx twice as fast as mbl_matrix_product_a_bt(AAt,A,A)

Definition at line 157 of file mbl_matrix_products.cxx.

void mbl_matrix_product_a_bt	(	vnl_matrix< double > &	ABt,
		const vnl_matrix< double > &	A,
		const vnl_matrix< double > &	B
	)

Compute product ABt = A * B.transpose().

NDEBUG

Definition at line 61 of file mbl_matrix_products.cxx.

void mbl_matrix_product_a_bt	(	vnl_matrix< double > &	ABt,
		const vnl_matrix< double > &	A,
		const vnl_matrix< double > &	B,
		int	nc
	)

Compute ABt = A * B.transpose() using only first n_cols columns of A and B.

Definition at line 80 of file mbl_matrix_products.cxx.

void mbl_matrix_product_adb	(	vnl_matrix< double > &	ADB,
		const vnl_matrix< double > &	A,
		const vnl_vector< double > &	d,
		const vnl_matrix< double > &	B
	)

ADB = A * D * B where D is diagonal with elements d.

A, d, and B should be compatible sizes, ADB will be resized to fit.

ADB = A * D * B where D is diagonal with elements d.

where D is diagonal with elements d

Definition at line 254 of file mbl_matrix_products.cxx.

void mbl_matrix_product_at_a	(	vnl_matrix< double > &	AtA,
		const vnl_matrix< double > &	A,
		unsigned	nc
	)

Compute product AtA = A.transpose() * A using nc cols of A.

Definition at line 211 of file mbl_matrix_products.cxx.

void mbl_matrix_product_at_a	(	vnl_matrix< double > &	AtA,
		const vnl_matrix< double > &	A
	)

Compute product AtA = A.transpose() * A.

Definition at line 246 of file mbl_matrix_products.cxx.

void mbl_matrix_product_at_b	(	vnl_matrix< double > &	AtB,
		const vnl_matrix< double > &	A,
		const vnl_matrix< double > &	B
	)

Compute product AtB = A.transpose() * B.

Result is of size A.cols() x B.cols()

Definition at line 112 of file mbl_matrix_products.cxx.

void mbl_matrix_product_at_b	(	vnl_matrix< double > &	AtB,
		const vnl_matrix< double > &	A,
		const vnl_matrix< double > &	B,
		int	nc_a
	)

Compute AtB = A.transpose() * B , using first ncols_a cols of A.

Result is of size ncols_a x B.cols()

Definition at line 164 of file mbl_matrix_products.cxx.