In page 50, in explaining Non-maximal Suppression, lecture note says NMS compares "intensity" of each pixel with two adjacent pixels.
But, in real NMS code, they compares "gradient magnitude".
I also think comparing "gradient magnitude" is right because NMS indirectly finds zero crossings, location of likely edges.
Then, in this page's explaining, does "intensity" mean "gradient magnitude" not "pixel value"?
In page 50, in explaining Non-maximal Suppression, lecture note says NMS compares "intensity" of each pixel with two adjacent pixels. But, in real NMS code, they compares "gradient magnitude". I also think comparing "gradient magnitude" is right because NMS indirectly finds zero crossings, location of likely edges. Then, in this page's explaining, does "intensity" mean "gradient magnitude" not "pixel value"?