Cross disciplinary biometric systems

More about the book

Cross-disciplinary biometric systems enhance the performance of conventional systems, significantly improving recognition accuracy and robustness in challenging environments like varying illumination. These advancements benefit face, fingerprint, iris recognition, and image search systems. Face recognition, for instance, is a natural and less intrusive method for identity verification, as people display their faces publicly. The effectiveness of face recognition systems is amplified through innovative ideas from color science, mathematics, and computer science, including pattern recognition, machine learning, and image processing. These ideas have led to the creation of new color models and effective color features in color science, innovative features derived from wavelets and statistics, and new kernel methods in mathematics. In computer science, advancements include new discriminant analysis frameworks, similarity measures, and image analysis methods that fuse multiple features from frequency, spatial, and color domains. Additionally, system design has evolved with new integration strategies and various fusion methods, such as feature-level and decision-level fusion, employing novel similarity measures to enhance overall performance.