Automatic Registration of OWAS Postures from Video Film

Abstract

The aim of this paper is to contribute to the development of a system for automatic registration of working postures from video film according to the OWAS method. The automatic system provides several advantages as compared with traditional manual systems. For the operator, it is a less demanding task and the analyses are more comprehensive. The system is a marker-less method based on image analysis and computer vision. Routines are developed both for separating the subject from its background and for relating the shapes of the extracted subject to OWAS postures. Several type of filters are used for the segmentation process and length and angle calculations on different segments of the extracted figure are executed to determine the corresponding OWAS posture. The determined angles can be used as input in biomechanical programs for calculating forces and load moments. Another approach for the development of the automatic system is to use neural network and feature matching. For that approach a neutral network is taught to relate postures of a subject to postures in the OWAS system.