AN INTEGRATED KFKT-LDA MECHANISM FOR FACIAL EMOTION EXTRACTION AND RECOGNITION WITH SVM AS CLASSIFIER

Major issue in the context of facial feature extraction is to extract the low dimensional facial feature. In this context Linear Discriminant Analysis (LDA) is the well known feature extraction mechanism intended for linear dimension reduction. In this paper, a novel mechanism for feature extraction is proposed based on FukunagaKoontz Transform (FKT). A kernel approach for FKT is proposed and integrated with LDA to improve the feature recognition rate.In addition Support Vector Machine (SVM) is used as a classifier for efficient emotion classification. The experimental results demonstrated in the article depicts that the proposed mechanism outperforms several existing mechanisms in the context of feature extraction and recognition.