contrib/mul/ipts/ipts_local_entropy.h File Reference

Compute entropy in region around each image pixel. More...

#include <vil/vil_image_view.h>

Go to the source code of this file.

Functions
void	ipts_local_entropy (const vil_image_view< vxl_byte > &image, vil_image_view< float > &entropy, unsigned h, unsigned min_v=0, unsigned max_v=255)
	Compute local entropy in square region around each pixel in image.

---

Detailed Description

Compute entropy in region around each image pixel.

Author:

Tim Cootes

Definition in file ipts_local_entropy.h.

---

Function Documentation

void ipts_local_entropy	(	const vil_image_view< vxl_byte > &	image,
		vil_image_view< float > &	entropy,
		unsigned	h,
		unsigned	min_v,
		unsigned	max_v
	)

Compute local entropy in square region around each pixel in image.

For each pixel in image, compute entropy in region (2h+1)x(2h+1) centred on the pixel. Result put in entropy image, which is of size (image.ni()-2h) x (image.nj()-2h). Thus entropy(i,j) corresponds to the value in the box around image point (i+h,j+h).

Values in image are assumed to lie in the range [min_v,max_v]. Any values outside that range will be ignored in the entropy calculation.

Local peaks are interesting points (see work of S.Gilles and that of T.Kadir).

For each pixel in image, compute entropy in region (2h+1)x(2h+1) centred on the pixel. Result put in entropy image, which is of size (image.ni()-2h) x (image.nj()-2h). Thus entropy(i,j) corresponds to the value in the box around image point (i+h,j+h).

Values in image are assumed to lie in the range [min_v,max_v]. Any values outside that range will be ignored in the entropy calculation.

Definition at line 29 of file ipts_local_entropy.cxx.