Computer Science Department Events

There will be a Colloquium on November 04 at 10:30 AM by Ketan Mayer-Patel

Friday November 04

E&CS Building Auditorium (1st Floor)

TIME: 10:30 AM

Title: Use of computer algorithms to discover novel regulatory elements of a plant biosynthetic pathway

Speaker: Ketan Mayer-Patel, UNC-Chapel Hill

Abstract:

About 10 years ago, my wife's aunts suffered a devastating fire from which they escape unharmed but lost everything they owned. By far the most precious of the items lost were the photographs and memorabilia that documented a life together. In the aftermath, friends and family went through their own collections in order to cull duplicates, find extra pictures from similar events, and scan originals in order to help restore their collection. The result did not recreate what they had exactly, but as an approximation, it comes close.

In this talk, I'll talk about ways in which we might think of multiple representations of digital media objects (primarily video) in which the same preservation benefit can be provided by independently owned and managed adaptive representations of the same original media object. We want to enable independent adaptation of media objects that maximize local utility while preserving all of the information of the original within the collective in a purely distributed and probabilistic manner. This would allow each owner of a particular media object to independently negotiate the trade off between quality and size as best benefits that owner and their use of the media object. Preservation of the original media object is embedded in the collective representations of all independent adaptations of the media object. In a nutshell, the greater the number of lower quality adaptations of a particular media object exist, the greater the likelihood that the original media object can be recovered. Furthermore, the burden of maintaining a "master" original media object is borne by no one in particular. The approach presented is theoretically applicable to a wide spectrum of existing video compression formats but highlights ways in which the structure of current formats hinders our ability to achieve this goal and suggests improvements and features for future formats.

Bio:

Dr. Ketan Mayer-Patel is an Associate Professor of Computer Science at the University of North Carolina, Chapel Hill. He received his BA, MS, and PhD, from the University of California, Berkeley in 1992, 1997, and 1999 respectively. Dr. Mayer-Patel's research interests are in multimedia computing and networking and his current projects focus on networking mechanisms to support tele-immersion. Dr. Mayer-Patel is the recipient of the NSF CAREER award and serves on the editorial boards of both ACM Transactions on Multimedia Computing, Communications, and Applications (TOMCCAP) and IEEE Multimedia Magazine. He is co-chair of the executive committee for the NOSSDAV workshop and a founding executive committee member for the conference ACM Multimedia Systems (MMSys).