00000ctmaa22000003a 4500 UP-1685675941123940355 Buklod 20160908102251.0 aa rb |||1 u| ta 160908s xx d r |||| u| (iLib)UPTAC-00018460477 UPVTC eng LG 993.5 2016 C66 C67 Cordero, Kenver R. author. HEXIG Hybrid Extreme Rotation Forest for offline signature verification Kenver R. Cordero; John D. Ultra, adviser. 2016. [11], 84 leaves illustrations. Undergraduate thesis (B.S. Computer Science) -- University of the Philippines, Tacloban. Nowadays, biometrics is widely used for personal identification and authentication systems because it is almost impossible to forge and it is unique for each person. Handwritten signatures which is a kind of biometrics, has been long established as the most widely-spread means of personal identification and has been commonly used in financial transactions and on establishments where security is of utmost importance. Since not every establishment can afford to implement a signature verification system, signature verification is done by some authority or clerical employee. As a result, signature verification has been a subjective and confusing process since every person's way of interpreting a signature is different and people don't always sign documents in the same manner. However, since the case where a fraudulent person replicates another person's signature to take his/her identity for illegal purposes is getting rampant, there is a need for establishments to implement a more effective signature verification system. This study proposes an automatic signature verification system using Hybrid Extreme Rotation Forest (HERF) for classifying genuine and forged signatures. Global, Mask and Grid features were extracted from the signature images and were used as inputs to the HERF ensemble. Results in this study shows that HERF is a very good algorithm to be used in offline signature verification. Signature verification. Ultra, John D. adviser. FI UP UPTAC UPTAC LG 993.5 2016 C66 C67 Thesis