A novel approach for fingerprint vitality identification with partial power of cosine transformation of fingerprint images based on machine learning techniques

Biometric fingerprint authentication is generally reliable and accurate, but it remains vulnerable to spoofing attacks using latent or fake fingerprints. This study addresses this challenge by proposing a novel fingerprint vitality recognition method that differentiates genuine fingerprints from latent or counterfeit ones. The method employs a machine learning paradigm across multiple stages: data collection, pre-processing, feature extraction, classification, and result analysis. Initially, an extensive dataset of fingerprint images is collected and pre-processed using Histogram Equalization to enhance brightness and reduce distortions, noise, and outliers, improving image quality for subsequent analysis. In the feature extraction stage, Partial Power of Cosine Transformation (PPCT) is applied to capture unique vitality-related signatures from the processed images. These features are then classified using the Gradient Descent Lion Pride Optimized XG-Boost (GDLP-XG-Boost) algorithm, a robust and effective classifier trained on annotated labels to distinguish genuine from spoof fingerprints accurately. The proposed approach is evaluated using standard performance metrics, including accuracy, Area Under Curve (AUC), sensitivity, and specificity. Experimental results demonstrate that the method effectively detects spoofed fingerprints while maintaining high classification performance, providing strong evidence of its reliability. This approach enhances the security of fingerprint-based biometric systems by offering a sophisticated solution against spoofing attacks. Overall, the study makes a significant contribution to improving the resilience and trustworthiness of fingerprint authentication systems, ensuring better protection against increasingly advanced and deceptive biometric fraud attempts.