Over 400 Attendees at ACM Multimedia 2006
Yahoo! researchers gathered last week for ACM Multimedia 2006, held October 23 to 27 in scenic Santa Barbara, California. Aside from sunshine, beaches and large fruity drinks donning plastic monkey straws, the conference covered themes of multimedia computing, including content analysis, processing and applications.
Certainly the biggest news was the increasing interest in Internet-scale multimedia databases, and human-centered multimedia processing. The world is filled with billions of images, millions of audio files and videos, and with billions of people searching for audio, images, and video for entertainment and information. Yahoo! researchers hope to understand it all by harnessing millions of Internet users as they find, use, share, and expand the wealth of multimedia data at their fingertips. This is the world of multimedia in 2006.
Yahoo!'s own Bradley Horowitz delivered one of two keynote addresses – a bold vision of the future where multimedia and communities come together to deliver a richer and more compelling experience, something that can be seen all over Yahoo!
Other contributions by Yahoo! Research included:
Patrick Schmitz was "Leveraging Community Annotations for Image Adaptation to Small Presentation Formats." He used the image annotations provided by flickr users to guide the presentation of an image on a small (mobile) device.
Patrick Schmitz, Peter Shafton, Ryan Shaw, Samantha Tripodi, Brian Williams, and Jeannie Yang demonstrated "International Remix: Video Editing for the Web," a Web-based application for editing video used for the San Francisco Film Festival.
Marc Davis delivered an invited paper in the Brave New Topics session on “Human-Centered Multimedia” about how large scale social media systems that connect content, context, and community will change the way media is produced and analyzed and also presented his vision of the future of multimedia computing in an invited panel on “Foundations and Directions of Multimedia Research.”
Ryan Shaw and Patrick Schmitz further described in "Community Annotation and Remix: A Research Platform and Pilot Deployment" how the different ways in which consumers are remixing video can help a system better reason about a video's content.
Kyogu Lee (Stanford) and Malcolm Slaney presented "Automatic Chord Recognition from Audio Using a Supervised HMM" in which they synthesized a very large database of music and used it to train a state-of-the-art chord-recognition system.
Malcolm Slaney and William White (Yahoo! Music) presented "Measuring Playlist Diversity for Recommendation Systems." They talked about measuring people's diverse musical interests, and how this can improve recommendation data.
SongHui Chon (Stanford), Malcolm Slaney, and Jonathan Berger
(Stanford) talked about "Predicting Success from Music Sales Data - a Statistical and Adaptive Approach." They used Billboard sales data to predict how long a new album will be on the charts.
Marc Davis and Patrick Schmitz spoke on a panel about “Multimedia and Web 2.0 - Hype, Challenge, Synergy” helping to challenge the community to think about how to better integrate Web 2.0 design ideas and multimedia platforms and applications research.
Mor Naaman, and Marc Davis talked about "Generating Summaries and Visualization for Large Collections of Geo-Referenced Photographs" in which they analyzed the Flickr photo collection, using geographic data to pick "representative" photos for different locations, and generating "Tag Maps" that visualize the photographic activity on a map.
Sengamedu Srinivasan presented two different demos.
