Real time estimation and tracking of human body Center of Mass using 2D video imaging

摘要

Estimating the position of the body Center of Mass (CoM) is a very important issue in human movement analysis in general, and in gait analysis as a more specified field. There are different methods to estimate this feature of gait. However, application of video imaging for analyzing human gait has increased tremendously due to its availability and non-interrupting nature. In this paper, a novel method is proposed for estimating the position of CoM using commercial VGA cameras. This method employs some of known algorithms in shape analysis, such as triangulation and using image signatures to segments body parts. Although the algorithms are well-known in the field of image processing. But their combination resulted in powerful segmentation method. Besides this segmentation method the proposed method benefits from a powerful background estimation method which makes it suitable to be used in real applications. The most important feature of the proposed method is its autonomy and speed, which make it suitable for use in sport analysis and elderly care. In addition to estimation of the CoM, this method can easily be applied to estimate other gait parameters, such as linear or angular speed of each link, joint angles. Such characteristics make the algorithm a suitable basis for building a complete gait analysis system.