Together with Yantao Zheng, Huanbo Luan, and Tat-Seng Chua from the National University of Singapore I am organizing a special session on ‘Challenges in Interactive Video Search’ at the forthcoming ACM CIVR 2008 conference. The session will include 5 invited presentations only, so it is not possible to submit to this session unfortunately.

Introduction
Due to advancements in video capturing and broadcasting technology, digital video is currently prosperously stored, shared, accessed, and distributed across Internet, digital libraries and other repositories. On a monthly basis, a media corporation might have thousand-hours of new broadcast and narrowcast video to maintain in their video database; while the video sharing site in the Internet, like YouTube, might have tons of new digital videos uploaded and accessed in their video repository. In spite of this upsurge in video data collection, present-day search technology is rather primitive and does not meet the requirements of real life users, nor provides it effective monitoring and management of large-scale video collections. The increasing demand for effective retrieval and indexing of digital video has spurted much research attentions of the growing video retrieval research community. Despite, the advancement in content-based video analysis, the divide between extractable low-level features and detectable high-level concepts on one hand versus user interpretation of video content on the other hand remains a gap too broad to bridge automatically. Thus, user interaction remains a critical component for effective and efficient video search.

Aim
Fueled by the recent success of evaluation campaigns, like TRECVID and the VideOlympics, the aim of the proposed special session is to highlight state-of-the-art scientific developments, innovative technologies, and major research opportunities in interactive video search. Besides effective solutions emphasizing retrieval performance, we will also highlight novel search methodologies that address the role of the expert user, the extension to novice users, the influence of collaborative interaction mechanisms, and advanced interfaces for efficient interaction. Specifically, the session aims to address the following challenges:

• How should relevance feedback and active learning techniques be incorporated effectively to maximize the effect of interaction between user and system?
• How to achieve both system scalability and real-time responsiveness?
• How should the user-interface leverage the benefits of having multiple query methods available for retrieval?
• What mechanisms should video search engines provide to facilitate novice users?

The presentations in this session are addressing precisely these important challenges of interactive video search.