This guest post on ACM Multimedia 2009 is by Gerald Friedland, one of our reviewers and a research scientist at the International Computer Science Institute in Berkeley, CA.
ACM Multimedia was held in Beijing, China this year. Throughout its 17 years of existence, ACM Multimedia has established itself as the top conference in the field. What distinguishes ACM Multimedia from other conferences is the mixture of hardcore technical content (the acceptance rate is about 15% or less for full papers) with very creative art and interactive video exhibitions, and a tradition of interesting competitions such as the open-source competition and the new ACM Multimedia Grand Challenge. To add to the diversity, the conference also features excellent tutorials, workshops, keynotes, panels, a doctoral symposium, and a brave new topics session. My summary covers only a small part of ACM Multimedia 2009, and of course it is biased by my personal viewpoint.
Technical Content
This year's conference full and short papers came from four different tracks. The content track deals mainly with content analysis algorithms, such as multimedia content indexing, retrieval, and recognition. The application track features multimedia tools and end systems and, together with the systems track, forms the more applied part of the conference. A recently introduced track is the human-centered track, which features work on the human-centered computing aspects of multimedia. Over the years, the balance of the tracks has changed slightly. In earlier years, systems and applications were weighted a little heavier. These days, most talks belong to the content track. Traditionally, most work in this track is image- and video-based. However, this year, one could observe a clear trend toward multimodality. In contrast to earlier years, I estimate more than about 10% of the contributions (papers, workshops, open-source competition) dealt with audio content analysis and a significant part of the conference dealt with true multimedia content analysis. It will be interesting to see if this trend continues in the future.
Here is a small selection of papers that I found interesting:
- Unfolding speaker clustering potential: a biomimetic approach by Thilo Stadelmann and Bernd Freisleben
- Changing timbre and phrase in existing musical performances as you like: manipulations of single part using harmonic and inharmonic models by Naoki Yasuraoka, Takehiro Abe, Katsutoshi Itoyama, Toru Takahashi, Tetsuya Ogata, and Hiroshi G. Okuno
- Short-term audio-visual atoms for generic video concept classification by Wei Jiang, Courtenay Cotton, Shih-Fu Chang, Dan Ellis, and Alexander Loui (this was also a best paper candidate)
- Comprehensive query-dependent fusion using regression-on-folksonomies: a case study of multimodal music search by Bingjun Zhang, Qiaoliang Xiang, Huanhuan Lu, Jialie Shen, and Ye Wang
- Interactive background blurring by Chih-Yu Yan, Ming-Chun Tien, and Ja-Ling Wu
- Personal photo album summarization by Pinaki Sinha, Hamed Pirsiavash, and Ramesh Jain
- MobileTI: a portable tele-immersive system by Wanmin Wu, Raoul Rivas, Ahsan Arefin, Shu Shi, Renata M. Sheppard, Bach D. Bui, and Klara Nahrstedt
Competitions
A major part of ACM Multimedia is competitions. Of course, it is prestigious to receive a best paper award at a conference with this acceptance rate. Therefore, the best paper award candidates (one from each track) are featured in one session. They were:
- Short-term audio-visual atoms for generic video concept classification by Wei Jiang, Courtenay Cotton, Shih-Fu Chang, Dan Ellis, and Alexander Loui
- Visual query suggestion by Zheng-Jun Zha, Linjun Yang, Tao Mei, Meng Wang, and Zengfu Wang (the winner)
- Design and deployment of a hybrid CDN-P2P system for live video streaming: experiences with LiveSky by Hao Yin, Xuening Liu, Tongyu Zhan, Vyas Sekar, Feng Qiu, Chuang Lin, and Hui Zhang, Bo Li
- Understanding near-duplicate videos: a user-centric approach by Mauro Cherubini, Rodrigo de Oliveira, Nuria Oliver
In the open-source competition, multimedia-relevant open-source projects compete for the title. This year, there were five finalists:
- Caliph & Emir: MPEG-7 photo annotation and retrieval by Mathias Lux (the winner)
- jReality: a java library for real-time interactive 3D graphics and audio by Steffen Weißmann, Charles Gunn, Peter Brinkmann, Tim Hoffmann, and Ulrich Pinkall
- xStreamer: modular multimedia streaming by Alexis Rombaut, Nicolas Staelens, Nick Vercammen, Brecht Vermeulen, Piet Demeester
- Music analysis, retrieval and synthesis of audio signals MARSYAS by George Tzanetakis
- TAPESTREA: a new way to design sound by Ananya Misra, Ge Wang, Perry R. Cook
For the first time this year, ACM Multimedia featured the Grand Challenge. It is a set of problems and issues proposed by industry leaders (such as Yahoo!, Google, HP, Nokia, and others), meant to engage the multimedia research community in solving relevant, interesting, and challenging questions about the industry’s two-to-five-year horizon for multimedia. The researchers were encouraged to submit working systems, as well as a short description in response to the challenge to win the Grand Challenge competition. The finalists were asked to pitch their idea and demo their system in exactly 180 seconds, with a 120-second question-and-answer phase. The selection of 14 finalists who, based on this format, presented top-notch ideas in rapid succession created a very colorful and fun event that was attended by virtually the entire conference audience of 450 people. More information about the event can be found here. The complete list of prizewinners can be found here.
Other Events
Larry Rowe was awarded the SIGMM technical achievement award, and gave a very interesting talk on the field of multimedia and making a career in it. Frank Nack hosted a very vivid interactive art program that was not to be missed.
The reception on the second day took place at the Beijing Film Academy. The film school is the largest institution specialized in the tertiary education for film and television production in Asia. The different showcases of both student work and the insight into the labs were not only exciting, but also a beautiful addition to the conference. The conference banquet was held in the historical Grand Ball Room of the Beijing hotel, and was one of the most impressive settings I have ever witnessed for a conference banquet.
In summary, this year's ACM Multimedia showed again that it is a conference that connects top-notch scientific work with fun and creative discussions, art, and show. I am looking forward to the coming years. ACM Multimedia 2010 is being held in Firenze, Italy.
Comments