A class to compute subpixel convolutions by using shifted operators. More...
#include <vil/algo/vil_correlate_1d.h>#include <vil/algo/vil_correlate_2d.h>#include <vil/vil_transpose.h>#include <vgl/vgl_point_2d.h>Go to the source code of this file.
Functions | |
| template<class srcT , class destT , class kernelT , class accumT > | |
| void | brip_subpix_convolve_2d (const vil_image_view< srcT > &src_im, vil_image_view< destT > &dest_im, kernelT kernel, accumT ac, int N) |
| Convolve an image(src_im) with shifted kernels to compute the value at the subpixel positions. | |
| template<class srcT , class kernelT , class accumT > | |
| void | brip_subpix_convolve_2d (const vil_image_view< srcT > &src_im, const vcl_vector< vgl_point_2d< accumT > > &pts, vcl_vector< accumT > &res, kernelT kernel, accumT ac, int) |
| Convolve the image with the given kernel at the given subpixel locations (pts) only. | |
| template<class srcT , class kernelT , class accumT > | |
| void | brip_subpix_convolve_2d (const vil_image_view< srcT > &src_im, const vcl_vector< vgl_point_2d< accumT > > &pts, const vcl_vector< accumT > &thetas, vcl_vector< accumT > &res, kernelT kernel, accumT ac, int) |
| Convolve the image with the given kernel at the given subpixel locations (pts) only at the given orientation. | |
| template<class srcT , class destT , class kernelT , class accumT > | |
| void | brip_subpix_convolve_2d_sep (const vil_image_view< srcT > &src_im, vil_image_view< destT > &dest_im, kernelT kernel, accumT ac, int N) |
| Convolve an image(src_im) with shifted kernels to compute the value at the subpixel positions. | |
A class to compute subpixel convolutions by using shifted operators.
Modifications
<none yet>
Definition in file brip_subpix_convolution.h.
| void brip_subpix_convolve_2d | ( | const vil_image_view< srcT > & | src_im, |
| vil_image_view< destT > & | dest_im, | ||
| kernelT | kernel, | ||
| accumT | ac, | ||
| int | N | ||
| ) | [inline] |
Convolve an image(src_im) with shifted kernels to compute the value at the subpixel positions.
The subpixel values are computed at 1/2^N intervals in each dimension. e.g., N=0: regular convolution, N=1: compute at mid-pixel locations, N=2: compute at quarter pixel locations
I'm using vil_correlate_2d_at_pt to evaluate the dot_product at pixel/subpixel locations dest_im is resized to [2^N*src_im.ni() x 2^N*src_im.nj()]
Definition at line 28 of file brip_subpix_convolution.h.
| void brip_subpix_convolve_2d | ( | const vil_image_view< srcT > & | src_im, |
| const vcl_vector< vgl_point_2d< accumT > > & | pts, | ||
| vcl_vector< accumT > & | res, | ||
| kernelT | kernel, | ||
| accumT | ac, | ||
| int | |||
| ) | [inline] |
Convolve the image with the given kernel at the given subpixel locations (pts) only.
Definition at line 93 of file brip_subpix_convolution.h.
| void brip_subpix_convolve_2d | ( | const vil_image_view< srcT > & | src_im, |
| const vcl_vector< vgl_point_2d< accumT > > & | pts, | ||
| const vcl_vector< accumT > & | thetas, | ||
| vcl_vector< accumT > & | res, | ||
| kernelT | kernel, | ||
| accumT | ac, | ||
| int | |||
| ) | [inline] |
Convolve the image with the given kernel at the given subpixel locations (pts) only at the given orientation.
Definition at line 139 of file brip_subpix_convolution.h.
| void brip_subpix_convolve_2d_sep | ( | const vil_image_view< srcT > & | src_im, |
| vil_image_view< destT > & | dest_im, | ||
| kernelT | kernel, | ||
| accumT | ac, | ||
| int | N | ||
| ) | [inline] |
Convolve an image(src_im) with shifted kernels to compute the value at the subpixel positions.
The subpixel values are computed at 1/2^N intervals in each dimension. e.g., N=0: regular convolution, N=1: compute at mid-pixel locations, N=2: compute at quarter pixel locations
To speed up the convolution, I'm performing 2 independent convolutions with 1-d kernels. dest_im is resized to [2^N*src_im.ni() x 2^N*src_im.nj()]
Definition at line 192 of file brip_subpix_convolution.h.
1.7.5.1