We argue how we should deal with some pieces of information each of which is not so strong for person recognition. On the basis of Dempster-Shafer theory, we introduce: 1) a new method of assigning a basic probability to nodes on a decision tree that is a basic expression of our current psychological status when we recieve an evidence, and 2) an update rule to combine several evidences presented sequentially. In person identification, the effectioness of these approaches is confirmed. © Springer-Verlag 2004.
CITATION STYLE
Yamada, M., & Kudo, M. (2004). Combination of weak evidences by D-S theory for person recognition. Lecture Notes in Computer Science (Including Subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), 3213, 1065–1071. https://doi.org/10.1007/978-3-540-30132-5_144
Mendeley helps you to discover research relevant for your work.