Other CMU DLI2 Related Projects
CMU dli2 project quick links:
SILVER Video Editor, Vivisimo, SAGE Project, Image Analysis and Understanding

The SILVER Video Editor

SILVER is a research project in the Human Computer Interaction Institute at Carnegie Mellon University that aims to enable easy video authoring by leveraging research conducted on the Informedia research project at Carnegie Mellon University. We will create a comprehensive Intelligent Video Editor that will allow people without special training to author interesting compositions using digital video. The Silver Project is funded in part by the National Science Foundation and others under Grant No. IIS-9817527 as part of the Digital Libraries Initiative Phase 2. Go to Project Abstract.

Project Director: Brad Myers, Senior Research Scientist, HCI Institute

Vivisimo

We are applying our Vivisimo document clustering algorithms and software to the challenge of spontaneously organizing Informedia video transcripts, noise and all. When a user searches for transcripts using a keyword search, our methods organize the search results hierarchically and display them using a split screen: browsable hierarchy on the left, and the results list on the right. The interaction style is based on Microsoft Explorer, and is novel for the task of document search/browsing.

Project Director: Raul Valdes-Perez, Senior Research Scientist, Computer Science Dept
Jerome Pesenti, Postdoctoral Visiting Scientist

The SAGE Project

The SAGE Project builds environmnets for expoloratory dta analysis through interactive information visualization. A video (21MB avi file or 7MB Real Media File) illustraties exploring metadata extracted by Informedia.

Project Director: Steve Roth , Senior Research Scientist, Robotics Institute
Mark Derthick, Project Scientist, Computer Science Department

Image Analysis and Understanding

In the CMU Robotics Institute we are conducting research in automatic image analysis. We are developing algorithms that can automatically recognize physical objects such as human faces and automobiles in photographs and video. These algorithms use statistical principles to represent visual appearance -- both of the object we wish to locate and of the general visual world. With our approach we have developed the most accurate face finder and car finder currently in existence. The links to related projects:
Face Detection and Object Recognition Using Statistical Modeling

Project Directors: Takeo Kanade, Director of the Robotics Institute
Henry Schneiderman, Research Scientist, Computer Science Dept