ARCHIVES

Multimodal biometric systems have emerged as a robust solution to overcome the limitations of unimodal approaches. This study presents a score-level fusion framework combining face and palm vein biometrics using Logistic Regression (LR) and Support Vector Machine (SVM) classifiers. The system employs Z-score normalization to standardize matcher outputs before fusion. Performance is evaluated using Equal Error Rate (EER) and Receiver Operating Characteristic (ROC) curves across five-fold cross-validation. Experimental results demonstrate that LR-based fusion achieves a lower mean EER (0.2116 ± 0.0229) compared to SVM-based fusion (0.2290 ± 0.0269). ROC analysis further confirms the superior performance and stability of LR across different operating points. The findings highlight the effectiveness of multimodal fusion and the suitability of linear probabilistic models for score-level integration.