Web proceedings papers


Ivana Atanasova and Ljupco Jordanovski


Biometric technology is often used for verification and identification of individuals and objects. Face recognition requires analyzing a particular face and feature processing. Amongst the earliest used methods for face recognition is the feature - based method. The local feature method first includes face region detection from still digital image, then component detection and finally features extraction. This paper develops the idea of a system which identifies a face based on a minimum number of features. Predefined number of features was extracted from the basic face components and analyzed over a training face database with variable expression images. Results from the test dataset confirmed that the minimum 7-feature set is suitable for proper face identification. Research also showed that family-related entities can be recognized and this is something yet to be validated.


face recognition, detection, feature-based, face components, family relations