Automatic recognition of facial expression using features of salient patches with SVM and ANN classifier
Visanthi V.P, Anu George · 2017 International Conference on Trends in Electronics and Informatics (ICEI) · 2017
Variety of methodologies has been proposed for facial expression recognition. Facial expression recognition has been very important in computer vision over the last several decades. This paper presents a novel method for facial expression detection using LBP features of selected salient patches on CK+ video Data Base. A few prominent active patches are extracted based on the position of facial landmarks. The positions of landmarks are obtained automatically. These active patches are further processed to obtain salient patches by 10 fold cross validation. Extracted features from the salient patches improve the computational cost with small variation in accuracy. Extracted features are classified by using both SVM and ANN classifier and results are compared.