A Comparison Between SURF and SIFT Methods For Biometric Feature Extraction
Abstract
The definition of biometric information can be presented as the usage of a certain trait, both physiological and biological, to calculate a person’s identity. One thing to note about biometric information is that it remains the same for one’s entire lifetime, and is different for each person (for example the iris, fingerprint etc...). In this paper, we have established an algorithm which uses the SURF (Sped-up Robust Features) in order to detect and extract data and has performance-scale- and rotation-invariant interest point detection and description. With this method it becomes possible to compute and make comparisons much faster, yet still is able to compete with, or even produce better results than previously proposed schemes SIFT (Scale Invariant Feature Transformation) concerning ease of repetition, uniqueness, as well as robustness. For this result to be gained, certain images are relied upon in order to undergo the convulsion process of the images. By identifying the areas of strength amongst the world’s best detectors and descriptors, (which is done with a Hessian matrix-based measure for the detector, and a distribution-based descriptor); SURF descriptors have been applied to object recognition and location, the recognition of people or faces, to reconstruct 3D scenes, to track objects and to extract points of interest. In this paper, we conclude that SURF is a powerful way to obtain accurate results at 93% and speed
Keywords
K-mean segmentation, Speeded Up Robust Features, SURF, Scale Invariant Feature Transformation, SIFTMetrics