AdaBoost-SVM Face Recognition Algorithm Based on LBP and Gabor Feature Fusion

In recent years, facial recognition technology has seen rapid advancements and is now extensively utilized in security surveillance and financial transactions. The face recognition process usually includes image preprocessing, feature extraction, and classification. However, it is often difficult for a single feature to fully capture the complex features of a human face, thus affecting the accuracy of recognition. To address this issue, this study presented an AdaBoost-SVM face recognition algorithm that fuses local binary pattern and Gabor features, which can enable comprehensive extraction of facial features from images and enhance recognition accuracy. This method first performs bilateral filtering on the face image to reduce noise and retain edge information. Then, it extracts the local binary pattern and Gabor features and weights the fusion. Finally, it uses the AdaBoost-SVM classifier for identification. Moreover, the algorithm performance was comprehensively evaluated using the ten-fold cross-validation method. The experimental results demonstrated that this method achieves recognition accuracy of $\mathbf{9 8. 7 5 \%}$ and $96.25 \%$ on ORL and Yale A face databases respectively, and still performs well in complex lighting and expression-changing environments, significantly improving the accuracy of face recognition and its robustness.