00000ctmaa22000003a 4500 UP-1685675941123939928 Buklod 20151116153432.0 aa rb |||1 u| ta 151116s xx d r |||| u| (iLib)UPTAC-00017910686 UPVTC eng LG 993.5 2007 C66 S36 Saño, French Andrew author. Offline fingerprint verification system French Andrew Saño, Russell Joann B. Suyom, [and] Miguel A Tonido, Jr.;John Paul T. Yusiong, adviser. 2007. [4], 58 leaves illustrations. Undergraduate thesis (B.S. Computer Science) -- University of the Philippines, Tacloban. Fingerprint is one of the most widely used biometric verification. Generally, a fingerprint verification system, takes advantage of the matching between acquired fingerprint image and the related database fingerprint image. A fast fingerprint verification method is based on the core and minutiae detection of the fingerprint. However, fingerprint images are of rarely perfect quality. They may be degraded and corrupted due to variations in skin and impression conditions. Thus, image enhancement techniques are employed prior to minutiae extraction to obtain a more reliable estimation of minutiae locations. The performance of minutiae extraction relies heavily on an enhancement algorithm. A bank of Gabor filters, oriented to different angles are applied to the image to clean it from noises that can result to false alarms or authentication mistakes. After the enhancement of the image, extraction of the core point using the poincare index algorithm is done to get the global features of the fingerprint. Then the extraction of the minutiae point follows. From the core, vectors will be traced to the minutiae for the purpose of image alignment and fingerprint matching. After extracting all the minutiae it will be matched to a template. The global and local fingerprint features will be the basis for the matching of the two fingerprints. It is based on the Euclidean distance between two corresponding feature vectors. If the Euclidean distance is within a threshold, then it is considered a match. In order to have a match between template and database fingerprint, it must reach a certain number of matched minutiae. To test the robustness, reliability and accuracy of the system it must have a low False Accept Rate and False Reject Rate. The system will be tested on the available fingerprint database in the net, the Fingerprint Verification Competition 2004 (FVC 2004). Biometric. Yusiong, John Paul T. adviser. Suyom, Russell Joann B. author. Tonido, Miguel Jr. author. FI UP UPTAC UPTAC LG 993.5 2007 C66 S36 Thesis