Semantic Video Annotation and Vague Query

The Digital Video Album (DVA) system described here integrates various cooperating subsystems to index and query video documents according to their semantic content and other metadata. A simple structured model is proposed to represent the video content. This model is compatible with XML Schemas and supports typed attributes and composition relationships. The architecture of DVA is described, and the workflows related to the semi-automatic indexing of videos are presented. The querying processing on video metadata is based on an extension of Boolean search, in a way to avoid empty answers and to rank query results. The query interface is also described.

• Type: Conference
• Year: 2003
• Where: MMM
• Authors: Qiuying Zhang, Mohan S. Kankanhalli, Philippe Mulhem
• Keywords: Digital Video Album, MMM 2003, Multimedia, Query Video Documents, Video