|
Paper title “Metrics Performance Evaluation: Application to Face Recognition”
Authors:
Naser Zaeri, Abeer AlSadeq, and Abdallah Cherri Abstract – Countless number of applications varying from music, document classification, image and video retrieval, require measuring similarity between the query and the corresponding class. To achieve this, features, that belong to these objects are extracted and modified to produce an Ndimensional feature vector. A database containing these feature vectors is constructed, allowing query vectors to be applied and the distance between these vectors and those stored in the database to be calculated. As such, the careful choice of suitable proximity measures is a crucial success factor in pattern classification. The evaluation presented in this paper aims at showing the best distance measure that can be used in visual retrieval and more specifically in the field of face recognition. There exist a number of commonly used distance or similarity measures, where we have tested and implemented eight of these metrics. These eight metrics are famous in the field of pattern recognition and are recommended by the Moving Picture Experts Group (MPEG). More than 300 tests on 300 different databases were performed to consolidate our conclusion. The evaluation shows that the Euclidean and Minkowski distance measures are the best. On the other hand, the Canberra distance measure gives the worst results.
|