I am having a comparison of different feature detection methods. So if we compare the SIFT (Scale-invariant feature transform) and Shi-Tomasi method, will their feature locations agree and why? Are they comparable? Thanks in advance.
$\begingroup$
$\endgroup$
$\begingroup$
$\endgroup$
Probably not. The SIFT detector finds centers of blob-like features. Shi-Tomasi detector finds corners. Furthermore, SIFT detector operates at multiple scales, while the classic Shi-Tomasi does not.
$\begingroup$
$\endgroup$
Depends. If you use two separate pre-canned libraries to compute them, likely not. However, note that when people talk about "SIFT features" they refer to two things:
- Point locations on the images
- Descriptors, a.k.a. collections of numbers computed from the pixels around the point locations.
What defines SIFT is really the descriptors, whereas the point locations can be computed any way you choose. A common choice is Harris corners, but nobody prevents you from using your method of choice - in particular, the Shi-Tomasi "best features to track" algorithm (I think this is the one you refer to).
