Biography | Publications

Mr. Swears has performed research in motion pattern learning, anomaly detection, track analysis and characterization using footage from building security cameras, aerial video, and various ground based sensors. He has also used solid state radar detections to research multi-object tracking, track linking, genealogy, and event detection.

After completing his BS at Rensselaer Polytechnic Institute in 2001, he joined the Discrimination Group at Lockheed Martin in Moorestown, NJ where he designed and developed algorithms for the Aegis system inter-continental ballistic missile tracker. In parallel, he earned his MS from Drexel University in Philadelphia, PA in early 2005 and progressed to the Tracker Team Lead in late 2005. In 2006, he joined the GE Global Research Center as a contractor, where he spent 18 months researching motion pattern learning and anomaly detection.

Mr. Swears has been employed at Kitware, Inc., since September 2007 as a member of the research staff with a focus on Computer Vision.  He has participated in several successful proposal efforts involving activity modeling and recognition including DARPA’s VIRAT (Video and Image Retrieval and Analysis Tool) project, on which he is an algorithm developer.  He will also contribute to VIRAT through his research at RPI under Professor Qiang Ji. 

Mr. Swears is the principal researcher working on CARVE (Complex Activity Recognition in Video), which is also a DARPA effort.  The goal of the project is to input a video clip of a football play, and determine which play from a known playbook is being executed, as early as possible, using combinations of weak evidence. He is also continuing to work on PANDA phase II, which is also a DARPA funded effort.  PANDA automatically learns normalcy models in an online unsupervised manner of open sea ship vessels while detecting anomalies.

Mr. Swears is pursuing his PhD at Rensselaer Polytechnic Institute (RPI) in Troy NY, where he is performing research under Professor Qiang Ji in the Intelligent Systems Lab. RPI is a Kitware subcontractor on the VIRAT program and Mr. Swears’ research at RPI will complement his research at Kitware.

Publications

E. Swears, A. Hoogs, and A. G. A. Perera. "Learning Motion Patterns in Surveillance Video using HMM Clustering" . Proceedings of the IEEE Workshop on Motion and Video Computing. Copper Mountain, CO. 9 Jan 2008.