I am working on a project on human-robot interaction that involves me capturing videos of two humans working on a task together. Eventually, I will analyze the videos and try to model how these humans are working together.
In order to capture their teamwork effectively, I set up three cameras at multiple angles (one at the starting location of the task, one at the ending location and an overhead camera that spans the entire task area). I have obtained these videos. Each of the cameras were started 1 second apart. This is because all three cameras are connected to a laptop computer, so I didn't want to overload the system with 3 camera worth of the video information at the same time.
The first step in modeling this, is to make sure that all the 3 videos are time-synced so that I can apply any number of modeling techniques that depend on time (such as HMM, CRF, DBN's etc). Furthermore, I would need to freeze frame at every time step (granularity of the time step would be determined later but might be as small as one-tenth of a second) and note different features for my modeling algorithm. I am a complete noob in video editing so if anyone could suggest a software that could accomplish what I desire, I would greatly appreciate it. I have access to all three platforms: Windows 7, Ubuntu 12.04, Mac OS X 10.7 (although I would prefer Windows or Ubuntu as my Macbook is quite old and might not be able to perform video editing with ease).
