I wanted to see if I understood this technique correctly- this is completely different from the Harris Corner detection technique right?
mpotoole
The point of this slide is that scale matters. For the left image, the curve within the smallest circle looks flat, and looks like a corner within the largest circle. For the right image, it's the opposite---it looks like a corner in the smallest circle, and looks flat(-ish) within the largest circle.
So when applying a Harris corner detector (or performing some other operation), the scale of the image / window size can affect your results. In addition to finding the x- and y- coordinates of a feature (e.g., corner), it is often important to identify the scale of the feature as well.
I wanted to see if I understood this technique correctly- this is completely different from the Harris Corner detection technique right?
The point of this slide is that scale matters. For the left image, the curve within the smallest circle looks flat, and looks like a corner within the largest circle. For the right image, it's the opposite---it looks like a corner in the smallest circle, and looks flat(-ish) within the largest circle.
So when applying a Harris corner detector (or performing some other operation), the scale of the image / window size can affect your results. In addition to finding the x- and y- coordinates of a feature (e.g., corner), it is often important to identify the scale of the feature as well.