The main paradigm that VideoQ operates is the visual one. Under the visual paradigm, the user can formulate a graphical query, sketching objects and adding features like color, texture and motion.

Each video clip has been annotated with 10-20 keywords. The text search engine is an effective way to search for a particular subject (sports), location (Paris), or object (lion).

The visual query is especially effective when combined with text search. The text search acts as a pre-filter that limits the number of initial videos to search. The visual query can further refine the search, locating specific details within a video shot.

Semantic Visual Templates represent an small set of visual queries that can effectively retrieve a category of videos. SVTs have been generated from a two-way interaction between the user and system.

A video taxonomy has been developed, and the video clips have been organized within this taxonomy. If you would like a quick overview of all the subject categories that are available, visit the Video Navigator.