"User-trainable video annotation using multimodal cues."

Ching-Yung Lin et al. (2003)
a service of  Schloss Dagstuhl - Leibniz Center for Informatics