Enhancing Local Binary Pattern-Support Vector Machine (LBP-SVM) for Improved Facial Emotion Classification

Supervised Machine Learning algorithms like Support Vector Machines (SVMs), have emerged as powerful tools for classification, regression, and anomaly detection tasks. However, despite their high potential for emotion classification, SVMs face several challenges, including manual hyperparameter tuning, overfitting, difficulties in handling overlapping classes and noisy features, and failure to effectively detect action units. Addressing these challenges is important in improving the accuracy and robustness of facial emotion recognition systems. This study proposed improvements to the Local Binary Pattern - Support Vector Machine (LBP-SVM) framework to address these challenges. The Grid Search will be used to automatically select the optimal value for C and gamma parameters. To address the problem of overfitting and noisy features, Recursive Feature Elimination will be used to select the features that contribute the most to the class separability. Finally, to further enhance the accuracy of the model the researcher also combined the feature extracted from Facial Action Units to capture the dynamic characteristics of facial expressions. Experimental results on the CK+ dataset show the enhanced LBP-SVM achieves 98% accuracy, demonstrating improved robustness for recognizing subtle and complex emotional expressions.