We describe two image matching techniques that owe their success to a combi
nation of geometric and photometric constraints. In the first, images are m
atched under similarity transformations by using local intensity invariants
and semi-local geometric constraints. In the second, 3D curves and lines a
re matched between images using epipolar geometry and local photometric con
straints. Both techniques are illustrated on real images.
We show that these two techniques may be combined and are complementary for
the application of image retrieval from an image database. Given a query i
mage, local intensity invariants are used to obtain a set of potential cand
idate matches from the database. This is very efficient as it is implemente
d as an indexing algorithm. Curve matching is then used to obtain a more si
gnificant ranking score. It is shown that for correctly retrieved images ma
ny curves are matched, whilst incorrect candidates obtain very low ranking.