H. Foroughi (Iran), A. Naseri (Canada), H.S. Yazdi, and H. Pourreza (Iran)
Human Motion Recognition, Video Surveillance, Fall Type Detection, Eigenspace, Spatio Temporal Database, Occurrence Stacking.
The development of intelligent video surveillance systems is so important due to providing secure and safe environments. To this end, this paper proposes an efficient novel approach for human motion recognition based on combination of integrated time motion images and eigenspace technique. Integrated Time Motion Image (ITMI) is a type of spatio-temporal database that includes motion and time of motion occurrence. Applying eigenspace technique to ITMIs leads in extracting eigen motion and finally a Multilayer Perceptron (MLP) neural network with back propagation learning scheme is used for precise classification of motions. Unlike existent human motion recognition systems that only deal with limited movement patterns, we considered wide range of motions consisting normal daily life activities such as walking, running, bending down, sitting down and lying down, some abnormal behaviors like limping or stumbling and also unusual events like falling. Reliable recognition rate of experimental results underlines satisfactory performance of our system. One of possible applications of the presented system is a fall incident detection system. While existing fall detection systems only can detect occurrence of fall behavior, the proposed system is able to detect type of fall event (Forward, Backward, Sideway) with a suitable average recognition rate.
Important Links:
Go Back
