Fake Fingerprint Classification Using Hybrid Features Learning With Gradient Boosting

Biometric security systems must be able to detect phony fingerprints to provide reliable authentication. The findings of this study suggest a hybrid approach to the detection of fake fingerprints that uses information on the texture and shape of the fingerprint. The novelty of this approach lies in...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:Applied Computational Intelligence and Soft Computing Jg. 2025; H. 1
Hauptverfasser: Ali, Muhammad Salman, Akram, Arslan, Rashid, Javed, Jaffar, Muhammad Arfan, Shah, Dilawar, Ali, Shujaat, Tahir, Muhammad
Format: Journal Article
Sprache:Englisch
Veröffentlicht: New York John Wiley & Sons, Inc 01.01.2025
Wiley
Schlagworte:
Access control
Accuracy
Biometric identification
Biometrics
Biometry
Classification
Cooperation
Deep learning
Fingerprints
Identification systems
Information systems
Literature reviews
Safety and security measures
Security systems
Spoofing
Wavelet transforms
Taiwan
ISSN:1687-9724, 1687-9732
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:Biometric security systems must be able to detect phony fingerprints to provide reliable authentication. The findings of this study suggest a hybrid approach to the detection of fake fingerprints that uses information on the texture and shape of the fingerprint. The novelty of this approach lies in integrating both traditional fingerprint information and geometric features obtained through wavelet transformation, which has not been extensively explored in previous studies. The proposed procedure uses the traditional fingerprint information and the geometric features that may be collected by wavelet modification. This allows it to take advantage of the complementary capabilities that these two types of capabilities offer. In addition, the hybrid feature set improves the system’s robustness and accuracy by leveraging each feature type’s unique strengths. To achieve this goal, the standard fingerprint information and the geometric aspects of the fingerprint are combined. It is possible to efficiently identify authentic and forged fingerprints by using these hybrid features and training a gradient boosting classifier. The findings of the studies demonstrate that the suggested technique achieves an accuracy of 96% on medium spoofing photos from the SOCOFing dataset, 97% on hard spoofing images, and 98% on mixed spoofing images. This high level of accuracy, especially on mixed spoofing images, showcases the effectiveness of the novel hybrid approach in diverse and challenging scenarios. This places it in the position of being the most accurate way currently accessible among the existing state‐of‐the‐art methods. Furthermore, the proposed method’s scalability and adaptability make it suitable for real‐world applications, potentially setting a new standard in biometric security. There is a great deal of optimism that the technique that has been described can increase the reliability and safety of biometric systems when used in situations representative of the actual world.
Bibliographie:ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 14
ISSN:1687-9724
1687-9732
DOI:10.1155/acis/8442143