PERSON IDENTIFICATION BASED ON MULTIMODAL BIOMETRIC RECOGNITION

• Main Author: ALEX, NG HO LIAN
• Format: Final Year Project Report
• Language: English; English
• Published: Universiti Malaysia Sarawak, (UNIMAS) 2020
• Subjects: TK Electrical engineering. Electronics Nuclear engineering
• Online Access: http://ir.unimas.my/id/eprint/37528/1/ALEX%20NG%20HO%20LIAN%2024pgs.pdf; http://ir.unimas.my/id/eprint/37528/2/ALEX%20NG%20HO%20LIAN%20ft.pdf; http://ir.unimas.my/id/eprint/37528/
• Institution: Universiti Malaysia Sarawak

Unimodal biometric systems have limited effectiveness in identifying people, mainly due to their susceptibility to changes in individual biometric features and presentation attacks. The identification of people using multimodal biometric systems attracts researchers' attention due to their advantages, such as greater recognition efficiency and greater security compared to the unimodal biometric system. A multimodal biometric system can overcome various unimodal biometric systems' limitations, so it is suitable and recommended use for this society. In this project, face and fingerprint recognition are used to develop a multimodal biometric system. In the process of face recognition, Classic Convolutional Neural Network (CNN) is used for training face datasets. After done training face dataset, the testing process is needed to recognize a face with face dataset. In the process of fingerprint recognition, the ORB algorithm is recommended to use in feature matching. ORB (Oriented FAST and Rotated BRIEF) algorithm consists of 3 stages: feature point extraction, defining feature point descriptors, and computing feature point matching. For these three stages, the fingerprint image is matching with the fingerprint database. For the process of fusion of face and fingerprint recognition, two features are fused by match score level fusion based on Weighted Sum-Rule. If the fusion score is higher than the threshold level is given, then the verification process is matched. The result of accuracy is displayed if the user selects the same biometric characteristics for both recognition. If the fusion score is less than the threshold level, then the verification process indicates a mismatch. The result of accuracy will not be displayed if the user selects different biometric characteristics for both recognition.