3D Front crawl swimming kinematics analysis: markerless vs marker based approach

Research on motion analysis of swimmers is commonly based on video recordings of the subject's motion, which are analyzed by manual digitization of feature points by an operator. This procedure has two main drawbacks: it is time-consuming, and it is affected by low repeatability. In order to overcome these drawbacks, the authors recently developed a video-based, markerless system for the analysis of arm movements during front crawl swimming (Ceseracciu et al. 2011). The aim of the present contribution is to evaluate accuracy of the proposed technique in estimating three-dimensional (3D) joint kinematics either in term of joints coordinates or angles. 3D coordinates of shoulder, elbow and wrist joints centers together with 3D elbow joints’ angles of an elite sprint swimmer, while performing front crawl swimming, were determined. Both markerless technique and the Calibrated Anatomical Systems Technique (CAST) (Cappozzo et al. 1995) were adopted. An automatic tracking technique was developed and employed in order to track the features necessary to apply the CAST technique. Root mean square distance (RMSD) values between trajectories and joint angles estimated with the two techniques were determined. Results showed good accuracy, and reliability (RMSD