Nathan Jacobs

research

My primary research area is in understanding long-sequences of video from outdoor cameras. I am working to solve traditional computer vision problems, such as calibration and tracking, using video of a scene captured over weeks, months, and even years.  This involves using natural scene variations to geo-calibrate widely distributed cameras and learning scene models to facilitate tracking and anomaly detection.  Please see my research overview for more details about this and other aspects of my work.

news

• 2013/02 paper on using cloud motion for camera calibration accepted to CVPR 2013
• 2013/01 article on depth from clouds accepted to PAMI
• 2013/01 article on using webcams to estimate cloud cover over the US presented at WACV 2013
• 2013/01 teaching Computer Vision
• 2011 selected as a participant in the 2011 DARPA CSSG program
• 2011 paper on analyzing small motions accepted to CVPR 2011

research highlights

	Shape from Clouds : In this work, we answer the question, "What can shadows cast by passing clouds tell us about 3D scene geometry?"
	Global Network of Outdoor Webcams : Tens of thousands of outdoor webcams are currently streaming live images from around the world. How can we use this imaging network to learn more about the world and how it changes over time?
	Inferring Subtle Motions : We are exploring methods of estimating very small but repeated scene motions by reusing the parameters of appearance-based representations. If you look closely at the video (left) you will see motion segmentation obtained using this method.
	Efficient Object Tracking : Object tracking can be made much more efficient by exploiting the consistent patterns of motion (e.g. cars tend to travel along roads) found in the scene. Here we explore an extreme variant of this that simply discards all image data not along the center line of a lane of traffic.

contact

The best way to contact me is probably through email at jacobs@cs.uky.edu.  For other choices please see my contact page.