Pablo Navarrete

Eigenspace-based face recognition: a comparative study of different approaches

Eigenspace-based face recognition corresponds to one of the most successful methodologies for the computational recognition of faces in digital images. Starting with the Eigenface-Algorithm, different eigenspace-based approaches for the recognition of faces have been proposed. They differ mostly in the kind of projection method used (standard, differential, or kernel eigenspace), in the projection algorithm employed, in the use of simple or differential images before/after projection, and in the similarity matching criterion or classification method employed. The aim of this paper is to present an independent comparative study among some of the main eigenspace-based approaches. We believe that carrying out independent studies is relevant, since comparisons are normally performed using the implementations of the research groups that have proposed each method, which does not consider completely equal working conditions for the algorithms. Very often, a contest between the abilities of the research groups rather than a comparison between methods is performed. This study considers theoretical aspects as well as simulations performed using the Yale Face Database, a database with few classes and several images per class, and FERET, a database with many classes and few images per class.

Eigenspace-based recognition of faces: comparisons and a new approach

Different eigenspace-based approaches have been proposed for the recognition of faces. They differ mostly in the kind of projection method used and in the similarity matching criterion employed. A first goal of this paper is to present a comparison between some of these different approaches. A second goal is to outline an adaptive, neural-based security access control system.

ANALYSIS AND COMPARISON OF EIGENSPACE-BASED FACE RECOGNITION APPROACHES

Different eigenspace-based approaches have been proposed for the recognition of faces. They differ mostly in the kind of projection method being used and in the similarity matching criterion employed. The aim of this paper is to present a comparative study between some of these different approaches. This study considers theoretical aspects as well as experiments performed using a face database with a few number of classes (Yale) and also with a large number of classes (FERET).

Interactive face retrieval using self-organizing maps

An interactive face retrieval system that uses self-organizing maps and user feedback is described. The system solves some problems of related content-based image retrieval systems: non-existence of trivial high-level human descriptions of the images and the gap between the high-level descriptions and the low-level features used to index the images.

FACERET: An Interactive Face Retrieval System Based on Self-Organizing Maps

The basic problem in content-based image retrieval is the gap between the high-level descriptions used by humans to describe image contents and the low-level features, such as color, texture and shape, used to automatically index images in databases. This problem is even harder when there are non-trivial high-level human descriptions of the images, as in the case of face images. In these cases the employment of user interaction in the retrieval process is a good possibility to solve this task. In this context, FACERET, a content-based face retrieval system that uses self-organizing maps and user feedback is here introduced. Some simulations of the FACERET operation are also shown.

Comparative Study between Different Eigenspace-Based Approaches for Face Recognition

Different eigenspace-based approaches have been proposed for the recognition of faces. They differ mostly in the kind of projection method been used and in the similarity matching criterion employed. The aim of this paper is to present a comparative study between some of these different approaches. This study considers theoretical aspects as well as simulations performed using a face database with a few number of classes.

Towards a Generalized Eigenspace-Based Face Recognition Framework

Eigenspace-based approaches (differential and standard) have shown to be efficient in order to deal with the problem of face recognition. Although differential approaches have a better performance, their computational complexity represents a serious drawback. To overcome that, a post-differential approach, which uses differences between reduced face vectors, is here proposed. The mentioned approaches are compared using the Yale and FERET databases. Finally, a generalized framework is also proposed.

Kernel-based Face Recognition by a Reformulation of Kernel Machines

Until now, linear methods like PCA and FLD have been widely tested for solving the problem of face recognition, showing good recognition rates. In this paper a general framework is going to be introduced in order to solve these problems using kernels, i.e. applying the same linear methods in high dimensional spaces. For this purpose a general solution of kernel machines is obtained in a different way to the current approaches. As a result of this process the problem of KFD, originally solved for two-class problems, is solved for an arbitrary number of classes, so that it becomes applicable for face recognition. Simulations are performed using a small face database (Yale Face Database) and a large face database (FERET).

On the Generalization of Kernel Machines

Taking advantage of the linear properties in high dimensional spaces, a general kind of kernel machines is formulated under a unified framework. These methods include KPCA, KFD and SVM. The theoretical framework will show a strong connection between KFD and SVM. The main practical result under the proposed framework is the solution of KFD for an arbitrary number of classes. The framework allows also the formulation of multiclass-SVM. The main goal of this article is focused in finding new solutions and not in the optimization of them.

Eigenspace-based Face Recognition

Eigenspace-based face recognition is a very well known and successful face recognition paradigm. Different eigenspace-based approaches have been proposed for the recognition of faces. They differ mostly in the kind of projection method been used and in the similarity matching criterion employed. The aim of this paper is to present a survey of these different approaches. A general framework of the eigenspace-based face recognition paradigm is presented and different approaches are compared using theoretical aspects and simulations performed using a face database.