This project is read-only.
A Kinect home security camera.

Security with Visual Understanding (SVU) is a hardware/software solution which provides a more accurate security camera. Typical security cameras cannot differentiate a person from a cat or a robot vacuum, nor can they differentiate the owner of the home from an ill-intentioned intruder. Since cats and trees are rarely responsible for theft, a security camera which can detect humans would greatly enhance security cameras’ utility.

SVU uses the Microsoft Kinect to provide these capabilities. SVU recognizes when a human enters the image and furthermore, is able to differentiate between known and unknown persons by maintaining a database of known persons’ skeleton dimensions. These combined capabilities allow SVU to deliver an intelligent, autonomous security system capable of detecting an unwanted intruder and alerting the owner of the intrusion, while minimizing false alarms.

Watch a video explanation of SVU. Read master's report on SVU.

SVU_flowdiagram.png


Read more about SVU in a scholarly paper written by Joseph Fluckiger. SVU was created as part of a Master's Report in fulfillment of a Master's degree in Software Engineering at University of Texas at Austin.

Last edited May 20, 2012 at 5:18 PM by josephflu, version 10