No abstract

We propose, in this paper, a new biometric identification approach which aims to improve recognition performances in identification systems. We aim to split the identity database into well separated partitions in order to simplify the identification task. In this paper we develop a face identification system and we use the reference algorithms of Eigenfaces and Fisherfaces in order to extract different features describing each identity. These features, which describe faces, are generally optimized to establish the required identity in a classical identification process. In this work, we develop a novel criterion to extract features used to partition the identity database. We develop database partitioning with clustering methods which split the gallery by bringing together identities which have similar features and separating dissimilar features in different bins. Pruning the most dissimilar bins from the query identity features allows us to improve the identification performances. We report results from the XM2VTS database.

Face recognition finds its place in a large number of applications. They occur in different contexts related to security, entertainment or Internet applications. Reliable face recognition is still a great challenge to computer vision and pattern recognition researchers, and new algorithms need to be evaluated on relevant databases. The publicly available IV 2 database allows monomodal and multimodal experiments using face data. Known variabilities, that are critical for the performance of the biometric systems (such as pose, expression, illumination and quality) are present. The face and subface data that are acquired in this database are: 2D audio-video talking-face sequences, 2D stereoscopic data acquired with two pairs of synchronized cameras, 3D facial data acquired with a laser scanner, and iris images acquired with a portable infrared camera.The IV 2 database is designed for monomodal and multimodal experiments. The quality of the acquired data is of great importance. Therefore as a first step, and in order to better evaluate the quality of the data, a first internal evaluation was conducted. Only a small amount of the total acquired data was used for this evaluation: 2D still images, 3D scans and iris images. First results show the interest of this database. In parallel to the research algorithms, open-source reference systems were also run for baseline comparisons.

No abstract