Sample Internship Description

Title: Video Gait Extraction

Host: David Alan Tahmoush

Division: ARL/SEDD - Sensors And Electron Devices Directorate

Location: Adelphi, MD


There are two complementary projects. One is taking videos with a person in them and extracting a skeleton using techniques like patch tracking, background subtraction, and change detection. This project has data and is ready to go. It is part of a larger project and your results will be used. The next step is determining what the person is doing based on past cases, where features like swinging arm motions are extracted from the video and related to people playing baseball or being in a fight. The other project uses LEDs that are placed around the body, then picked up by cameras and extracting a moving body model. This could be the ground truth to the video approach, but we have not collected high-quality data yet. I can try to get some of this data if you are interested. We have some lower quality data already, and some systems to get more. The animation is taking that motion captured data or some of my earlier skeletons and animating them with a human skin trapezoidal mesh to produce a moving human. We have the movie-making software Maya to help, or you can build a mesh animation. The moving mesh would then help with matching simulations to measured data. You could also work on ray-tracing with the moving mesh models if you are interested. There is a lot to work going on here with good resources and the potential to publish papers, get good recommendations for graduate school, learn new skills, and improve your resume.