New Biometric Recognition System By Multichannel Orthogonal Radial Triangular Pulse Moments

Biometrics involves measuring biological characteristics to identify or authenticate individuals based on specific traits. A biometric system functions by capturing biometric data, extracting relevant features, and comparing these with stored data in a database. Depending on the medium used, biometric systems can focus on various characteristics such as iris patterns, facial features, fingerprints, or signatures. These systems employ various feature extraction techniques to facilitate recognition. This study presents a performance analysis of reconstruction and classification using K-Nearest Neighbors with $k=1$ (KNN), Support Vector Machines (SVM), and multichannel orthogonal radial triangular pulse moment invariants. Additionally, the performance of the proposed invariant feature vectors for biometric recognition is evaluated by numerical experimental experiments. Finally, by comparing our methods with real-time static characteristics, we demonstrate their potential efficacy in accurately recognizing patterns, highlighting their applicability within the field of biometric systems.