Currently, I'm focusing on two aspects of video streaming. First, aspect, which is also the topic of my PhD thesis, can be summarized as "finding appropriate video quality for computer vision". More and more streaming applications use video analysis algorithms to automate their tasks. These applications include video surveillance, video conferencing, automated mobile systems (robots, unmanned vehicles that rely on video analysis), social web-based applications, etc. I study the problem of saving network bandwidth when, instead of humans, video analysis algorithms observe the video or images. This problem is a part of research project DAVIS.
Another area I am interested in is the problem of arranging multiple live and recorded video streams on one screen. This problem is important in presentation capturing and video conferencing. I actively participated in the project Plasma, a scripting language and tool, which allows combining several video streams into one similarly to arranging GUI widgets on the screen. Another related project I did during my internship at FXPAL, where I have developed a web-based system to monitor and control presentation capturing process. The system relies on a hardware solution that can compose two live video feeds from a projector and a camera into a single MPEG stream.
My full research statement can be downloaded
here [PDF], 4 pages
and one-page version here [PDF].
Video with overview presentation of my research is available
here.
Also, the video presentation of my project at FXPAL can be viewed
here.