Feature Extraction Methods for Real-Time Face Detection and Classification
- PDF / 1,378,353 Bytes
- 11 Pages / 600 x 792 pts Page_size
- 55 Downloads / 253 Views
Feature Extraction Methods for Real-Time Face Detection and Classification David Masip ´ Centre de Visi´o per Computador (CVC), Departamento de Informatica, Universitat Aut`onoma de Barcelona, Bellaterra 08193, Spain Email: [email protected]
Marco Bressan ´ Centre de Visi´o per Computador (CVC), Departamento de Informatica, Universitat Aut`onoma de Barcelona, Bellaterra 08193, Spain Email: [email protected]
Jordi Vitria` ´ Centre de Visi´o per Computador (CVC), Departamento de Informatica, Universitat Aut`onoma de Barcelona, Bellaterra 08193, Spain Email: [email protected] Received 22 December 2003; Revised 29 November 2004 We propose a complete scheme for face detection and recognition. We have used a Bayesian classifier for face detection and a nearest neighbor approach for face classification. To improve the performance of the classifier, a feature extraction algorithm based on a modified nonparametric discriminant analysis has also been implemented. The complete scheme has been tested in a real-time environment achieving encouraging results. We also show a new boosting scheme based on adapting the features to the misclassified examples, achieving also interesting results. Keywords and phrases: face detection, face recognition, boosting, feature extraction.
1.
INTRODUCTION
As computers become faster and faster, new applications dealing with human faces become possible. Examples of this applications are face recognition applied to surveillance systems, gesture analysis applied to user-friendly interfaces, or gender recognition applied to reactive marketing. We will propose here a global face detection and recognition framework, which has achieved good results in an uncontrolled environment. Usually, working under uncontrolled conditions is one of the hardest problems of computer vision, for example, in applications where illumination presents strong changes, or where we have to deal with objects under unpredictable movements. We have tested the system in a real environment, with no restrictions on scale and illumination, achieving real-time satisfying performance. According to [1] face detection schemes can be classified into four different categories, although some methods can belong to more than one category. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
(1) Knowledge-based methods, where some rules or relationships between features are encoded. Kotropoulos and Pitas followed this approach using projection profiles to locate the face [2]. We have used a similar technique to locate the eyes once the face has been detected using our method. (2) Feature-invariant approaches, where the idea is to detect the facial features first, such as eyes, mouth, eye brows, and group them into candidate faces [3]. (3) Template-matching methods, where there is a predefined face pattern that is correlated with the image. Point distribution models (PDM
Data Loading...
