A Review on Facial Emotion Recognition Techniques

As a major phase skin Recognition, together with elliptical boundary model, is accomplished. Further, facial Feature Identification process is carried out. The next step is to initiate a technique for extorting geometric and anthropometric facial characteristics. At last we train as well as test the classifiers. We accomplished a categorization precision of 58.6% for six types of emotions (bliss, anguish, curiosity, despair, fury, hatred) and mean efficiency of 96.8% for two emotions (bliss and curiosity).The current study utilizes interest points as markers in face images that are damaged by few emotions as well as correlates its location to that of a normal expression. The outputs are viewed in contrast with Paul Ekman’s FACS (Facial Action Coding System) tool to check on the efficacy of the algorithm .The automated identification of face expressions utilizing image template matching method faces various issues pertaining to facial features and recording circumstances. Although this field has reached great heights , choosing of features as well as categorization method for emotion identification , till today remains an unsolved mystery. To suppress feature outliers, the proposed technique comprises of pixel normalization which is used to eliminate intensity offsets backed up using a Min-Max metric in a nearest neighbor classifier. The proposed Min-Max classification technique has an efficiency of 92.85% to 98.57% when checked on JAFFE database. Classification task also done using KNN, SVM and Bagged Tree Classifier.