Component based face recognition system

Abstract

The goal of this research paper was to design and use a component based approach to face recognition and show that this technique gives us recognition rates of up to 92%. A novel graphical user interface was also developed as part of the research to showcase and control the process of face detection and component extraction and to display the recognition results.The paper essentially consists of two parts, face detection and face recognition. The face detection system takes a given image as the input from which a face is located and extracted using 2D Haar Wavelets and Support Vector Machines. The face region is then used to locate and extract the individual components of the face such as eyes, eyebrows, lips and nose which are then sent to the face recognition system where the individual components are recognized by using Wavelets, Principal Component Analysis and Error Backpropagation Neural Networks. Pattern dimension reduction technique is used to significantly reduce the dimensionality and complexity of the task. © Springer Science+Business Media B.V. 2008.