Automated Facial Expression Detection using Machine Learning Algorithm

The main aim of the project is to find the novel feature extraction technique along with the suitable machine learning algorithm to detect the facial expression from the face image. This project mainly deals to overcome the drawbacks of existing feature extraction along with machine learning algorithm applied for facial image to detect the emotions. The scope of the project is to do image classification with lesser image by applying machine learning models rather than Deep learning methodologies, whereas the proposed technique can be applied to any face related classification when there is lack of images. In this paper, I have implemented a novel feature extraction method for face image which is the combination of both wavelet and geometrical features to overcome the existing drawbacks of stand-alone features. Further, those features are trained with different classifiers such as Support Vector Machine (SVM), K-Nearest Neighbour (K-NN) and Single Hidden Layer Feed Forward Neural Network (SHLFNN), and found that Multi-class SVM stands better for the extracted features. The evaluation method has carried out with different metrics such as Accuracy, Sensitivity, Specificity, F-Score and Kappa Statistic to support the proposed methodology