A Wavelet-Based Memory Autoencoder for Noncontact Fingerprint Presentation Attack Detection

Fingerprint presentation attack detection (FPAD) is essential in fingerprint identification systems. Noncontact methods such as fingerprint biometrics are becoming popular because they are not affected by skin conditions and there are no hygiene issues. However, most of the existing noncontact FPAD...

Full description

Saved in:
Bibliographic Details
Published in:IEEE transactions on information forensics and security Vol. 19; pp. 8717 - 8730
Main Authors: Liu, Yi-Peng, Yu, Hangtao, Fang, Haonan, Li, Zhanqing, Chen, Peng, Liang, Ronghua
Format: Journal Article
Language:English
Published: IEEE 2024
Subjects:
anomaly detection
autoencoder
Biological system modeling
biometrics
Feature extraction
Fingerprint recognition
frequency
Frequency-domain analysis
Image matching
Image reconstruction
memory
Memory modules
noncontact fingerprint
Presentation attack detection
wavelet
ISSN:1556-6013, 1556-6021
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Fingerprint presentation attack detection (FPAD) is essential in fingerprint identification systems. Noncontact methods such as fingerprint biometrics are becoming popular because they are not affected by skin conditions and there are no hygiene issues. However, most of the existing noncontact FPAD methods are supervised methods with poor generalizability and poor performance during events such as unseen presentation attacks (PAs). Moreover, easily overlooked frequency domain information contributes to the fingerprint antispoofing task. Therefore, we propose a wavelet-based memory-augmented autoencoder that fully utilizes the frequency domain information. Specifically, the model first decomposes the input image into high- and low-frequency information and extracts features separately. Subsequently, we propose a frequency complementary connection (FCC) module to realize the fusion and complementation of frequency domain information at the feature level. Moreover, a memory distance expansion loss is proposed to keep the memory module diverse. Experiments are conducted to verify the effectiveness of the method. The code of our model is available on https://github.com/SuperIOyht/WaveMemAE .
ISSN:1556-6013
1556-6021
DOI:10.1109/TIFS.2024.3463957