Abstract
The video surveillance system is used in a various fields, such as crime prevention, manufacturing industry etc. Conventional video surveillance system can detect abnormal motion from only TV. However, it is difficult to detect abnormal motion from only image automatically by computer. Therefore, we propose a monitoring system using image and depth data from Microsoft's Kinect sensor. We adopted higher order local autocorrelation (HLAC) features and cubic higher order local autocorrelation (CHLAC) features for detecting abnormal motion from image and depth data.
