Prof. Bernd Girod, Stanford University
Handheld mobile devices, such as camera phones or PDAs, are expected to become ubiquitous platforms forvisual search and mobile augmented reality applications. For mobile image matching, a visual data baseis typically stored at a server in the network. Hence, for a visual comparison, information must beeither uploaded from the mobile to the server, or downloaded from the server to the mobile. With relativelyslow wireless links, the response time of the system critically depends on how much information must be transferred in both directions. We review recent advances in mobile matching, using a "bag-of-visual-words" approach with robust feature descriptors, and show that dramatic speed-ups and power savings are possible by considering recognition andcompression jointly. We will use real-time implementations for different example applications, such asrecognition of landmarks or CD cover, to show the benefit from image processing on the phone, the server,and/or both.
Bernd Girod is Professor of Electrical Engineering and (by courtesy) Computer Science in the InformationSystems Laboratory of Stanford University, California. He was Chaired Professor of Telecommunications in the Electrical Engineering Department of the University of Erlangen-Nuremberg until 1999. His research interests are in the areas of video compression, networked media systems, and image databases. He has published over 450 conference and journal papers,as well as 5 books. Professor Girod has been involved in several startup ventures, among them Polycom (Nasdaq:PLCM), Vivo Software, 8x8 (Nasdaq: EGHT), and RealNetworks (Nasdaq: RNWK). He received the Engineering Doctorate from University of Hannover, Germany, and an M.S. Degree from Georgia Institute of Technology. Prof. Girod is a Fellow of the IEEE and of EURASIP and a member of the German National Academy of Sciences. He received the 2002 EURASIP Best Paper Award, the 2004 EURASIP Technical Achievement Award, and the 2007 IEEE Multimedia Communication Best Paper Award.
Dr. Roelof van Zwol, Yahoo! Research
Finding similar and relevant media content given a user query or sample image has been at the core of the multimedia retrieval community for a long time. In this talk, I will identify and address multimedia challenges that play a role at Yahoo!, and which go beyond relevancy of images and video to a given multimedia retrieval task. The true challenge for multimedia is to find a balance between relevancy, freshness, quality, interestingness and diversity in order to provide an engaging rich media experience to the user.
Roelof van Zwol is a senior research scientist at Yahoo! Research, where he is managing the multimedia research team. He has more than 10 years of international research experience in multimedia, information retrieval, (social) media mining, object ranking, spatial search, XML, databases, and machine learning. He is passionate about conducting research in an industrial context and to apply the outcomes in high-impact end-user services. His work on object ranking now powers the left-rail search suggestions in Yahoo!'s Web and image search engine.Prior to joining Yahoo he was an assistant professor at Utrecht University in the Netherlands. He received his Ph.D. in Computer Science in 2002 from the University of Twente in the Netherlands. Roelof van Zwol is the author of more than 70 peer reviewed publications, a number of which appeared in first-tier conferences such as SIGIR, WWW, and ACM Multimedia. He is an active member of the research community, who serves on the PC of conferences such as SIGIR, WWW, ACM Multimedia, CIVR, and ICMR. He was co-chair of CIVR in 2009, and organizer of the second CHORUS conference on Multimedia Search, as well as the organizer of 5 workshops themed around Web search, multimedia, and information retrieval. In European context, he was the technical coordinator of the Semedia EU project on search environment on Media, and project manager in the WeKnowIt EU project.
ICMR2011 - ACM International Conference on Multimedia Retrieval