Text Box: Department of Computer Science at Cornell University
Text Box: Text Box: The Salton Series is supported by Amit Singhal, Cornell PhD ‘97




Text Box: Although photographs don't lie, they often fail to capture our own experience of a scene. For instance, photographs may show closed eyes or facial expressions we would not notice in real life. Photographs rarely are able to capture the grandeur of canyons, mountains, and cityscapes. Blue skies come out white and shadows black. In this talk, I will first discuss what I mean by a "moment" and try to convince you that the "moment" is what we intend to capture when we point a camera at a scene. After a brief overview of newly developed software tools in computer vision, I will then present a body of recent work in Computational Photography that leverages these tools to overcome many of the limitations of traditional photographs. This will range from composing multiple images to create seamless Groupshots, to capturing high dynamic range imagery, to creating multi-gigapixel images to capture and display even the most grand landscapes.

This talk also represents a sort of homecoming, after embarking in the field of computer graphics 26 years ago at Cornell. I hope to shed a bit of light on the joys of following the twisted road of research that Cornell set me on.
Text Box:                The      
                Lecture Series           
Text Box: Thursday
October 15, 2009
Text Box: 4:15 pm
B17 Upson Hall
Reception - 4th Floor Atrium at 3:45pm

Michael Cohen

Principle Researcher—Interactive Visual Media Group


Michael F. Cohen joined Microsoft Research in 1994 from Princeton University where he served on the faculty of Computer Science. Before, Dr. Cohen served on the Architecture faculty at Cornell University and was adjunct faculty at the University of Utah.  Dr. Cohen received his Ph.D. in 1992 from the University of Utah. He also holds undergraduate degrees in Art (Beloit College) and Civil Engineering (Rutgers University), and an M.S. in Computer Graphics (Cornell).


His early work at Cornell and Princeton focused on the radiosity method for realistic image synthesis, for which he received The 1998 SIGGRAPH Computer Graphics Achievement Award. At Microsoft, Dr. Cohen has worked on a number of projects, including computational photography, image based rendering, animation, camera control, and more artistic, non-photorealistic rendering. The animation work focused on means to allow simulated creatures to portray their emotional state. In the 1990s, his foundational work on image-based rendering, dubbed The Lumigraph, captured the complete flow of light from an object for later rendering from arbitrary vantage points. Recent work has focused on computational photography applications, ranging from segmentation and matting of images and video, technologies for combining a set of images as a Photomontage, to the creation and viewing of high dynamic range, very high resolution panoramas.

Text Box: Moments: 
Why Cameras Fail to Capture Them and What We Can Do About It

Gerard Salton (1927- 1995) A towering figure in the field of information retrieval, Gerard Salton synthesized ideas from mathematics, statistics, and natural language processing to create a scientific basis for extracting semantics from word frequency. The impact of his contributions is profound - five textbooks, over 150 research papers, and dozens of Ph.D. students. The modern computer science and information science research scene, with its terabyte databases, Web, and related technologies, owes a great deal to Gerry's pioneering efforts.


This lecture series honors our former colleague with speakers who similarly are innovators in their fields.