Video Browser Showdown by NUS

Jin Yuan, Huan-Bo Luan, Dejun Hou, Han Zhang, Yan-Tao Zheng, Zheng-Jun Zha, Tat-Seng Chua

Abstract: The known item search task (KIS) aims to retrieve a unique video or video clip in the video corpus. This paper presents a novel interactive video browsing system for KIS task. Our system integrates visual content-based, text-based and concept-based search approaches. It allows users to flexibly choose the search approaches. Moreover, two novel feedback schemes are employed: first, users can specify the temporal order in visual and conceptual inputs; second, users can label related samples with respect to visual, textual and conceptual features. Adopting these two feedback schemes greatly enhances search performance.