Computer Science | Grinnell College

At noon on Friday, October 9, Dr. Harold Trease of the Pacific Northwest National Laboratory will speak on Video analytics for indexing, summarization and searching streaming video and video archives:

Given streaming video or video archives, how does one effectively summarize, classify, and search the information contained within such a large amount of image data? In this presentation, we address these issues by describing a process for the automated generation of a table of contents and of keyword, topic-based index tables that can be used to catalogue, summarize, and search large amounts of video data. Having the ability to index and search the information contained within the videos, beyond just metadata tags, provides a mechanism to extract and identify useful content. During this presentation, we describe some of the mathematics, computer science and engineering, and applications of being able to use image and video keywords as the primary search criteria, much as Web browsers (such as Google) allow us to search text today.

Dr. Trease is a graduate of the University of Nebraska at Kearney and received his doctorate from the University of Illinois at Urbana-Champaign, in nuclear engineering. He has more than thirty years of research experience in the design, implementation, and application of high-throughput, high-performance computer software. He currently leads the P3D Code Development Project. P3D is a large-scale framework for modeling, simulation, and prediction in computational physics.

Pizza and soda will be served before the talk. Everyone is welcome to attend!

This lecture serves as this week's CS Table.

On Thursday, October 8, Alex Exarhos 2010 will present the results of his summer research on analyzing videos:

I worked with a set of video indexing and searching algorithms through a Department of Homeland Security internship. These algorithms are capable of quickly providing a summary of a video by breaking it up into sections based on the content, and this method of indexing makes it possible to instantly locate other images, frames, or video segments within the indexed library. I will also talk about the web services I created for these algorithms, and the work I did implementing a mobile interface for a Google Android phone.

Refreshments will be served at 4:15 p.m. in the Computer Science Commons (Noyce 3817). The talk, Interfaces for video analytics, will follow at 4:30 p.m. in Noyce 3821.