Authors

G.P. Hegde – Professor, Dept. of Information Science and Engg., SDMIT, Ujire, Mangalore
gphegde@sdmit.in

Ashwini B – Assistant Professor, Dept. of Information Science and Engg., SDMIT, Ujire, Mangalore
ashwinib@sdmit.in

Abstract

Improvement of facial expression recognition under various occlusions for different applications is a challenging task. This paper emphasizes fusion of holistic color and texture features of face images, reducing feature vector dimensions using subspace methods like PCA, SVD, ICA, and FFA to improve classification accuracy.

Facial expression recognition under different critical conditions is complex and requires robust solutions that integrate machine learning and subspace methods to enhance accuracy.

Keywords

Facial Expressions, Feature Extraction, Subspace Methods, PCA, SVD, ICA, FFA, SVM Classifier

Conclusion

This work demonstrates the effectiveness of subspace methods in improving facial expression recognition under various occlusion scenarios. Feature fusion techniques significantly enhance recognition rates compared to conventional approaches.