The concept of Content Recommender system within the TV production is being developed within LIVE project to address the issues of the on-the-fly content selection from the TV archives during live production and personalization of the TV channels according to the preferences of the target audience groups.

The production of multiple stream TV program such as live sport transmission requires large amounts of source content. While multiple live streams from different cameras or different venues (in case of Olympic Games) can be combined to fill out several output streams, this might not be enough to produce targeted sub-channels containing more in-depth stories such as sport event background information, athlete interviews, or historical content. Such channels require massive reuse of vast audiovisual information stored in the TV archives.

The main goal of Content Recommender system is to enable usage of archive content within live production scenario. It provides an user interface for the professional user through which he is able to search and review archive clips. On the other hand, it provides automatic methods which combine preferences of the target audience (the TV viewers) with the predefined topics of the TV program to find relevant archive content in the archive. In this way it supports the production of personalized and targeted TV channels, directed to specific TV audience groups.