1 Fast Min-hashing Indexing and Robust Spatio- temporal Matching for Detecting Video Copies CHIH-YI CHIU, National Chiayi University HSIN-MIN WANG, AND CHU-SONG CHEN Institute of Information Science, Academia Sinica ________________________________________________________________________ The increase in the number of video copies, both legal and illegal, has become a major problem in the multime- dia and Internet era. In this paper, we propose a novel method for detecting various video copies in a video sequence. To achieve fast and robust detection, the method fully integrates several components, namely the min-hashing signature to compactly represent a video sequence, the spatio-temporal matching scheme to accu- rately evaluate the video similarity compiled from the spatial and temporal aspects, and some speed-up tech- niques to expedite both min-hashing indexing and spatio-temporal matching. The results of experiments demon- strate that, compared to several baseline methods with different feature descriptors and matching schemes, the proposed method that combines both global and local feature descriptors yields the best performance when encountering a variety of video transformation. The method is very fast, requiring approximately 0.06 seconds to search for copies of a thirty-second video clip in a six-hour video sequence. Categories and Subject Descriptors: H.3.3 [Information Storage and Retrieval]: Information Search and Retrieval - Information filtering, Search process; I.2.10 [Artificial Intelligence]: Vision and Scene Understand- ing - Video analysis General Terms: Algorithms, Design, Experimentation, Performance Additional Key Words and Phrases: Content-based copy detection, near-duplicate, histogram pruning ________________________________________________________________________ 1. INTRODUCTION With the rapid development of multimedia technologies, digital videos are now ubiqui- tous on the Internet. According to a report by AccuStream iMedia Research (http://www.accustreamresearch.com ), in 2006, the quantity of video streams increased 38.8% to 24.92 billion in media sites world-wide. One of the most popular video sharing sites, YouTube, hosted about 6.1 million videos, and 65,000 video clips were uploaded everyday. The enormous growth in the amount of video data has led to the requirement for efficient and effective techniques of video indexing and retrieval. In particular, since digital videos can be easily duplicated, edited, and disseminated, video copying has be- come an increasingly serious problem. A video copy detection technique would thus be helpful for protecting and managing video content. For instance, with such a technique, content providers could track particular videos with respect to royalty payments and Some parts of this work were published in Chiu et al. [2007]. This work was supported in part by the National Science Council of Taiwan under Grants NSC 98-2218-E-415- 003 and NSC 99-2631-H-001-020. Authors' addresses: C. Y. Chiu, National Chiayi University; email: cychiu@mail.ncyu.edu.tw; H. M. Wang, and C. S. Chen: Institute of Information Science, Academia Sinica, 128 Academia Road, Section 2, Nankang, Taipei 115, Taiwan. © 2008 ACM 1073-0516/01/0300-0034 $5.00