Behavior Classification Method Based on Skeleton Model from Video Images

Now the demand for intelligent video processing is getting greater and greater, moving body behavior classification from video images is their focus and difficulty. An approach to extracting human body and classifying the behaviors of the moving objects is presented in this paper. A statistical Gaussian model is used as adaptive background updating method. After foreground objects are segmented from the video images, we propose a novel skeleton model of the foreground objects model for feature extraction. The angles that can represent the pose of the skeleton model and length-width ratio of the human are used as feature vector. Then Bayesian Networks is used as classifier for behavior classification. The experiments show good recognition rate for moving human behavior classification.