SAIC Logo
Georgia Tech College of Computing Logo

Camera Sensor Network / System Virtualization

Faculty Advisor, Umakishore Ramachandra, PhD

Phase I-Camera Sensor Network:

The project is about video surveillance using multiple cameras.The Camera sensor network is a bigger project and has a broader perspective. The specific problem which I will be working on this semester is "How to support a real time face recognition system for a large scale camera network". Face recognition is usually noisy and very resource intensive. Doing that in real time for multiple faces in multiple camera may seem to be a very hard task. But use of context sensitive information and prioritization can make it tractable. One approach might be to do profiling of the various resources and algorithms available and depending on the time available use one or more of the algorithms /resources at hand. There might be other approaches and we are still looking for the right one. This is a really interesting problem and has lot of systems issues to be taken care of.

References:

1.(Junsuk Shin, Rajnish Kumar, Dushmanta Mohapatra, Umakishore Ramachandran, Mostafa Ammar) ASAP: A camera sensor network for situation awareness. (To Appear in OPODIS 2007)

2.(P. Kulkarni, D.Ganeshan, P.Shenoy, Q.Lu) Senseye: A multi-tier Camera Sensor Network (MULTIMEDIA'05)

3.(J.Krumm, S.Harris,B.Meyers,B. Brumitt, M. Hale, S. Shafer) Multi-camera multi person tracking for easy living (IEEE International Workshop on Visual Surveillance 2000)

4.Java Media Framework API

5.(S.L. Dockstader, A.M.Tekalp) Multiple camera fusion for multi-object tracking (WoMOT 2001)

6.Face Recognition Home Page (http://www.face-rec.org/)

Phase II- System Virtualization:

This project is in the virtualization domain.

Virtualization means the underlying hardware being virtualized by some Hypervisor sitting on top of hardware. In this way multiple OSes can run simultaneously on the same machine (each having their own virtual machine and thus isolated form each other). Xen and VMWare frameworks are examples of Virtualized systems.

Virtualization brings many exciting ideas to the world of computer science. Migration of Virtual machines is one of them which helps in migrating whole computing environment from one host machine to another.

Also the present day mobile devices like cell phones and Flash Disks have enough disk capacity to store the status of an entire virtual machine and transfer it from one machine to another and get started in the new machine with the same system(/software) environment as in the old machine. This is a really powerful concept and has lots of implication to the present day world. The exact problem I will be working on in this area is related to storage (disks/file systems) in such kind of scenarios. I am still in the stage of formulating the problem/research idea.

Camera Sensor Network: The project is about video surveillance using multiple cameras.The Camera sensor network is a bigger project and has a broader perspective. The specific problem which I will be working on this semester is "How to support a real time face recognition system for a large scale camera network". Face recognition is usually nosiy and very resource intensive. Doing that in real time for multiple faces in multiple camera may seem to be a very hard task. But use of context sensitive information and prioritization can make it tractable. One approach might be to do profiling of the various resources and algortihms available and depending on the time avaialble use one or more of the algorithms /resources at hand. There might be other approaches and we are still looking for the right one. This is a really interesting problem and has lot of systems issues to be taken care of.

References:

1.(I. Pratt et al.) Xen and the Art of Virtualization (SOSP 2003)

2.(C. Clark, K.Fraser, S.Hand, J.G.Hansen, E.Jul, C.Limpach, I.Pratt, A.Warfield) Live Migration of Virtual Machines (NSDI 2005)

3.VMWare (http://www.vmware.com)

4.VMWare Virtual Center (http://www.vmware.com/products/vi/vc)

5.(Fred Douglis et al.)Storage Alternatives for Mobile Computers (OSDI 94)

6.(Muthukaruppan Annamalai, Andrew Birrel, Dennis Fetterly, Ted Wobber) Implementing Portable Desktops: A New Option and Comparisons (Microsoft Research Technical Report)