Abstract

Recently, interest point detectors and descriptors have become prominent in the field of computer vision and are typically used to determine correspondences between two images of the same scene. The finite element scale invariant detector (FESID) is based on a similar multi-scale approach to that used in the SURF detector. However, FESID detects point features rather than blob features, and by combining the derivative and smoothing operations into a single operator efficient performance is achieved. We illustrate the performance of the FESID algorithm with respect to both robustness and correct region matching.