2008

REU Participants:

• Joshua Hartman, University of Florida
• Nicholas Hirsch, Rice University
• Lam Tran, University of Rochester
• Karthik Prabhakar, Indiana University-Purdue
• Ajay Chabra, University of Southern California
• Alexis Oyama, University of Nevada-Reno
• Robertson Augustine, University of Central Florida
• Steven Braeger, University of Central Florida
• Sarah Applegate, University of Central Florida
• Jason Hochreiter, University of Central Florida
• Cynthia Atherton, University of Central Florida

Joshua Hartman (University of Florida) worked on algorithms for creating high-resolution, sharp images from blurred, low-resolution images. He is currently working on extending this work and writing a report.

Lam Tran (University of Rochester) worked on incorporating machine-learning into shape-from-shading models. This work led to a system that can learn the optimal way to take advantage of visual cues about an object's shape.

Karthir Prabhakar (Indiana University-Purdue University) worked on recognizing activities from aerial video. He achieved improved results over algorithms published in the computer vision literature.

Jason Hochreiter(UCF) worked on developing detectors for recognizing actions in video sequence, specifically actions involving picking up and putting down objects. This work was incorporated into UCF's entry into the 2008 TRECVID video retrieval competition.

Ajay Chabra (New York Institute of Technology) worked on developing detectors for recognizing the movements of individuals in crowds. This work was also incorporated into UCF's entry into the 2008 TRECVID video retrieval competition.

Alexis Oyama (University of Nevada - Reno) worked on detecting chaotic motion in crowds. He put together a dataset of ground-truthed examples and used techniques from flow analysis to look for chaotic actions.

Nicholas Hirsch (Rice University) worked on separating moving objects from stationary objects. This has application to robotic platforms, which must find moving objects when planning motion.

Cynthia Atherton (UCF) worked on detecting hands and fingers using a new type of feature.

Sarah Applegate (UCF) worked on creating programming demos that use computer vision and graphics to demonstrate programming concepts to high-school students. This work has been intregated into an NSF-sponsored program for attracting high-school students into STEM disciplines.

Robertson Augustine (UCF) worked on statistical models of pedestrian movements in a scene. He both created new, more sophisticated models and created more efficient implementations of existing models.

Steven Braeger(UCF) developed a new algorithm for estimating a camera transfer function from a single image. This work uses more accurate camera models than previously used.