Menu Search & Directory

Computer Science Colloquium

Friday, 01 November, 2019

The Computer Science Colloquium will host Guna Seetharaman, Ph.D., senior scientist for advanced computing concepts with the U.S. Navy and chief scientist for the Naval Research Lab’s Center for Computational Studies, as its guest lecturer at 3:10 p.m. on Friday, Nov. 1, in Room 115, Ritter Hall.

Seetharaman will present, “Progress in Computer Vision: On Board Processing and Machine Perception.”

Lecture Abstract:

The explosive growth in data derived from video surveillance systems, has been driven by progress in sensor technologies, ubiquity of the mobile sensors, and ever increasing set of applications that exploit video data. Basic assumptions that were at the heart of various computer vision algorithms, no longer apply.  Moreover, the computing power accessible in-situ a camera has grown several orders of magnitude over the last two decades.

These have led to several paradigm shifting opportunities for performing machine perception in many real-time applications with acceptable latencies.  The talk will highlight a set of recent milestones, in video registration, onboard computation of 3D point clouds, robust tracking overcoming parallax errors, and suggest new directions in computing architecture for machine perception.  Examples will be drawn from Wide Area Motion Imagery (WAMI), and autonomous sensor networks comprised of airborne video systems. 

Applications for remote sensing, machine assisted farming, disaster relief operations etc., will be discussed in addition to battlefield awareness, including the  role of high performance computing at the edge.

About the speaker:

Seetharaman is the Navy senior scientist for advanced computing concepts, and the chief scientist of the Information Technology Division in the Naval Research Laboratory. He also serves as a senior scientific Advisor at the C5ISR-EW Directorate, Office of the Undersecretary of Defense Research and Engineering.  

He has worked on computer vision, parallel computing and machine perception algorithms, for over thirty years.