Finger Vein Recognition Systems: A Review on Presentation Attack Detection

Abstract- Finger vein recognition is an essential biometric identification technique that has become popular for security
enhancement in recent years. This paper presents various features sets examined in finger vein-based identification and
verification systems. The system comprises of image acquisition, preprocessing, feature extraction, and decision-making.
The different types of image acquisition techniques are light transmission, light reflection, bottom lighting, and side
lighting. The preprocessing techniques are further classified into quality evaluation, area of interest (ROI) selection,
normalization, and enhancement. It modifies the acquired images for better analysis. The feature extraction includes vein
pattern analysis, local pattern recognition, dimensionality reduction, and texture-shape descriptors to improve
identification. The acquired image is matched with the template stored in the database in the final stage. This paper also
presents various challenges in finger vein recognition systems, but it primarily focuses on the challenges in presentation
attack detection (PAD), that reduces the efficiency of the biometric systems. The limitations of PAD using federal learning
are also discussed.