Abstract
Facial expression recognition has been acknowledged as an active research topic in computer vision community. The challenges include the face identification and recognition, suitable data representation, appropriate classification scheme, appropriate database, among others. In this paper, a new approach for facial emotion recognition is investigated. The proposal involves the use of Haar transform and adaptive AdaBoost algorithm for face identification and Principal Component Analysis (PCA) in conjunction with minimum distance classifier for face recognition. Two approaches have been investigated for facial expression recognition. The former relies on the use of PCA and K-nearest neighbour (KNN) classification algorithm, while the latter advocates the use of Negative Matrix Factorization (NMF) and KNN algorithms. The proposal was tested and validated using Taiwanese and Indian face databases.
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call