Functions to compute histogram of orientations (HOGs) More...
#include <vil/vil_image_view.h>Go to the source code of this file.
Functions | |
| template<class srcT , class sumT > | |
| void | mipa_orientation_histogram (const vil_image_view< srcT > &src, vil_image_view< sumT > &hog_image, unsigned n_angles, unsigned cell_size, bool full360=true, bool bilin_interp=true) |
| Generate an image containing histograms of oriented gradients (HOG). | |
Functions to compute histogram of orientations (HOGs)
Definition in file mipa_orientation_histogram.h.
| void mipa_orientation_histogram | ( | const vil_image_view< srcT > & | src, |
| vil_image_view< sumT > & | hog_image, | ||
| unsigned | n_angles, | ||
| unsigned | cell_size, | ||
| bool | full360, | ||
| bool | bilin_interp | ||
| ) |
Generate an image containing histograms of oriented gradients (HOG).
At each pixel in src, compute angle and quantise into n_angles. If full360, then angle range is 0-360, else it is 0-180. hog_image is set to have n_angles planes. hog_image(i,j,k) gives the weighted sum of pixels with angle k in cell (i,j), corresponding to the i,j-th cell_size square block.
The corner of cell(0,0) is at src(1,1), to ignore border pixels.
Number of cells (size of hog_image) chosen so every cell entirely within src. Thus hog_image.ni()=(src.ni()-2)/cell_size. If bilin_interp=true then histogram contributions are added to neighbouring bins with weights inversely related to distance to centre of bin
At each pixel in src, compute angle and quantise into n_angles. If full360, then angle range is 0-360, else it is 0-180. hog_image is set to have n_angles planes. hog_image(i,j,k) gives the weighted sum of pixels with angle k in cell (i,j), corresponding to the i,j-th cell_size square block.
The corner of cell(0,0) is at src(1,1), to ignore border pixels.
Number of cells (size of hog_image) chosen so every cell entirely within src. Thus hog_image.ni()=(src.ni()-2)/cell_size.
Definition at line 25 of file mipa_orientation_histogram.txx.
1.7.5.1