Master of Science in Media Arts and Sciences, May 1995.

Previous research in digital video orchestration has demonstrated the computer's ability to edit sequences based upon a cooperation between video databases and encoded story knowledge. These systems, however, manipulate sound and picture as a single unit. The next generation digital video systems will need to combine the encoded story knowledge with the ability to manipulate audio and video separately. This thesis creates the infrastructure to support computer orchestrated editing in which sound and picture are treated as separate, complementary tracks.

MILESTONE is comprised of two modules. A video annotation component which builds upone previous research in video databases provides the means to create a knowledge represenation. A robust set of editing functions provide editing functionality. Potential applications of and arguments for similar applications to MILESTONE are discussed in light of theoretical inquiries into hypertext, cinematic narration, and video orchestration.

Read the thesis in HTML form