Local feature-based image matching: robustness to intra class variation, affine transformation and background clutter
Lixin Fan, Gabriela Csurka, Chris Dance
We present a feature-based image matching algorithm which establishes correspondences between individual
objects of the same class (e.g. face) in two images. The algorithm is robust to intra-class variation as well as
significant scale and view angle change, and background clutter. Excellent matching results are obtained
under these conditions.
ICPR 2004 Workshop Learning for Adaptable Visual Systems Cambridge, United Kingdom 22 August, 2004.