INTRODUCTION: Although underwater swimmers video recordings are widely diffused, swimmers kinematics is usually reconstructed only on a single sagittal view. Some authors performed a 2D kinematics analysis of the front crawl by means of multi-cameras underwater acquisition systems, relying on manual digitization of specific features` position. Thus in this study an automatic and repeatable method to assess 3D kinematics of front crawl swimming is presented. This combines both the Calibrated Anatomical Systems Techniques technique (CAST) [1] and a semi-automatic feature tracking technology. METHODS: 4 out of several front crawl trials performed by an elite swimmer and recorded by means of 5 underwater color analog cameras were selected. The latter were synchronized automatically by means of an ad-hoc application, and calibrated with Bouguet method. A set of 25 features was drawn on the body, to reconstruct the 3D kinematics of the right trunk and arm. The 3D trajectories of the following 4 anatomical landmarks were reconstructed by applying CAST: the incisura jugularis (IJ), the acromion (ACR), the center of the elbow joint (EJ) and of the wrist joint (WJ). The features were tracked in each video sequence by using an application based upon the registration technique proposed by Lucas and Kanade [2]. A comparison with the manual tracking technique, assessed by two trained experts, was also performed. RESULTS: The 3D joint trajectories repeatability analysis for CAST reported intra-method correlation coefficients between trials ranging from 0.936 to 0.999 and from 0.801 to 0.996 for EJ and WJ, respectively, along the 3 axes. A good correlation was found also for the IJ and the ACR (0.947 to 0.999 and 0.766 to 0.999) between trials along medio-lateral and antero-posterior axis, however along the vertical axis a poor correlation was found, especially for the ACR. Without applying CAST, the trajectories of the IJ and the ACR were determined during 10% of the acquired motion, reaching the 80% when CAST was applied. Further, the comparison with manual tracking showed good consistence of the proposed technique in terms of variability. DISCUSSION: The use of a combined methodology which employs both the semi-automatic tracking and CAST provides more objective and consistent results than direct manual video digitization. REFERENCES: 1. A. Cappozzo et al. Clinical Biomech, 4 171-178, 1995 2. B.Lucas,
© Copyright 2010 Biomechanics and Medicine in Swimming XI - Abstracts. Veröffentlicht von Norwegian School of Sport Sciences. Alle Rechte vorbehalten.