00000ctmaa22000003a 4500 UP-1685675941123958217 Buklod 20180817104123.0 aa rb |||1 u| ta 180817s xx d r |||| u| (iLib)UPTAC-00019010337 UPVTC eng LG 993.5 2018 C66 J33 Jadulco, Sushmita S. HERFing offline fingerprint verification system using Hybrid Extreme Rotation Forest. Sushmita S. Jadulco 2018 xiii, 110 leaves illustrations. A special problem (B.S. Computer Science) -- University of the Philippines, Tacloban. With the increasing need of technology, and with the world becoming more digitized over the years, protecting confidential information has become more challenging. Ease of access to information had led to a great need of security against network breaches and identity thefts. The advancement of biometric authentication methods has continued to bring solutions to this ever-changing problem. Fingerprint based security systems have become widely popular due to their potential to reliably identify people based on their near-universal physical trait. Its widespread use on different applications has shown its superiority over traditional methods of authentication. This study proposes an automatic fingerprint verification system using an innovative ensemble learning algorithm called Hybrid Extreme Rotation Forest (HERF) for classifying valid and invalid fingerprints. The features extracted from the fingerprint images includes global features, and the number of minutiae endings and bifurcations found on four regions of the fingerprint. Gabor filter-based fngerprint image enhancement was also used in this study to improve the result of the feature extraction. Results of this study show that HERF is a reliable and effective algorithm to be used in offline fingerprint verification systems. Ultra, John D. adviser. FI UP UPTAC UPTAC LG 993.5 2018 C66 J33 Thesis