Slap Fingerprint Segmentation Evaluation II - Procedures and Results
Craig I. Watson
In 2004, NIST conducted a fingerprint slap segmentation study  to assess the state-of-the-art in fingerprint segmentation technology. Given the development of new technology it has become necessary to reassess the current state-of-the-art of segmentation algorithms. SlapSegII will give providers of this technology the opportunity to participate multiple times as their technology improves and compare their results to previous results on a fixed standard database. The SlapSegII testing strategy, evaluation data, and measure of successful segmentation are discussed in detail in this evaluation testing plan. The results from SlapSegII are very encouraging. Many of the segmentation algorithms tested were able to successfully segment at least 3 fingers from a 2 or 3 inch slap image for 96-99% images (both right and left hand). Not surprisingly, the crease (bottom of the fingerprint) proved the most difficult edge to correctly detect. All segmentation algorithms had overall improvement if tolerances were relaxed or ignored around the crease.