Abstract
In this work we present an appearance-based 3-D Face Recognition approach that is able to recognize faces in video sequences, independent from face pose. For this we combine eigen light-fields with probabilistic propagation over time for evidence integration. Eigen light-fields allow us to build an appearance based 3-D model of an object; probabilistic methods for evidence integration are attractive in this context as they allow a systematic handling of uncertainty and an elegant way for fusing temporal information. Experiments demonstrate the effectiveness of our approach. We tested this approach successfully on more than 20 testing sequences, with 74 different individuals. © Springer-Verlag Berlin Heidelberg 2002.
Cite
CITATION STYLE
Krüger, V., Gross, R., & Baker, S. (2002). Appearance-based 3-D face recognition from video. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 2449 LNCS, pp. 566–574). Springer Verlag. https://doi.org/10.1007/3-540-45783-6_68
Register to see more suggestions
Mendeley helps you to discover research relevant for your work.
