Robust Fall Detection Using Human Shape and Multi-class Support Vector Machine

Falls and resulting physical-psychological consequences in the elderly are a major health hazard and a serious obstacle for independent living. So development of intelligent video surveillance systems is so important due to providing safe and secure environments. To this end, this paper proposes a novel approach for human fall detection based on human shape variation. Combination of best-fit approximated ellipse around the human body, projection histograms of the segmented silhouette and temporal changes of head pose, would provide a useful cue for detection different behaviors. Extracted feature vectors are finally fed to a multi-class support vector machine for precise classification of motions and determination of a fall event. Unlike existent fall detection systems that only deal with limited movement patterns, we considered wide range of motions consisting of normal daily life activities, abnormal behaviors and also unusual events. Reliable recognition rate of experimental results underlines satisfactory performance of our system.