Setting up the patch
In lines 34-36 we begin the basic setup of the gabor patch by calling setup_gabor. Because the
gabor is drawn on a per-pixel basis, it would take far to long to calculate all pixel values and to blit them individually.
Rather, only a quarter of the values are calculated and then they are rotated and copied for the other 3/4 of the gabor patch.
We declare that the width (grid_w) and height (grid_h) of one quarter of the
gabor patch is 100 pixels. windowsd refers to the standard deviation.
setup_gabor makes a call to bivariate_normpdf(), which is the bivariate
normal probability density function used to determine the gaussian blur effect.
Previous || 1 2 3 4 5 6 || Next