<![CDATA[Fingerprint image identification plays a crucial role in biometric authentication systems; however, selecting an appropriate algorithm remains challenging due to trade-offs between identification accuracy and computational efficiency. This study aims to comparatively evaluate the performance of a traditional Minutiae-Based algorithm and a Convolutional Neural Network (CNN) for fingerprint image identification to determine their respective strengths and limitations. The Minutiae-Based method extracts distinctive ridge features, such as ridge endings and bifurcations, followed by a similarity-based matching process. In contrast, the CNN model automatically learns discriminative features from raw fingerprint images through deep learning. Experiments were conducted on a multi-subject fingerprint dataset, and performance was assessed using identification accuracy, False Acceptance Rate (FAR), False Rejection Rate (FRR), and computation time per image. The results show that CNN achieved a higher identification accuracy of 98.6%, with a FAR of 1.2% and FRR of 1.4%, outperforming the Minutiae-Based algorithm, which obtained 92.3% accuracy, 4.8% FAR, and 3.9% FRR. However, the Minutiae-Based approach demonstrated superior computational efficiency, requiring an average processing time of 0.42 seconds per image compared to 1.35 seconds for CNN. These findings highlight a clear performance trade-off between accuracy and processing speed. The novelty of this study lies in providing a structured quantitative comparison that integrates accuracy, security metrics, and computational cost within a unified evaluation framework. The results contribute to the development of biometric systems by offering practical guidance for selecting fingerprint identification algorithms based on application-specific requirements, whether prioritizing high recognition accuracy or real-time computational efficiency.]]>
