INTRODUCTION: Since 2016, the agility test [1] has been used in performance diagnostic tests and in the return-to-sport process of the German alpine skiing, ski cross and ski freestyle team athletes. While the test has proven in practice to be very effective in the return-to-sport process, the assessment of both the two landings and the 15 jumps was subjective, with three categories each: good, average and bad. The subjective assessment is to be expanded, objectified and, ideally, replaced by an AI-supported, automated kinematic evaluation (Nemo, Simi Reality Motion Systems GmbH ©, Unterschleißheim, Germany). METHODS: All attempts were filmed frontally by only one camera (2D). The video recordings made, went through the automated AI process and as a result the joint points were transformed to pixel coordinates. The quality of the landings was subjectively assessed analogously to Stensrud [2]: Performance was reported concerning lateral tilt of the pelvis (not significant - some - obvious), Valgus motion of the knee (no obvious - slightly - clearly) and medial/ lateral side-to-side movements of the knee during the performance (no - some - clear). For comparison, the position of the hip axis in relation to the shoulder axis (hip to shoulder) and to the horizontal (hip to horizontal), the leg axes in relation to each other and the knee movement during landing were calculated from the coordinates. To assess the jumps, the determined flight and contact times, 2D center of gravity in flight and contact times and the horizontal distance from the middle of the foot to the middle of the knee (distance foot to knee) during the contacts were used. The subjective assessment had the criteria of position of the foot and leg axes, stable body position and overall coordination. RESULTS/DISCUSSION: The parameters determined showed a wide range in the results, greater than a rigid 3-point assessment. The quality of the valgus movement determination in the knee was sufficient. As of now, all results that were related to spatial expansion were more difficult to assess and therefore significantly less satisfactory. CONCLUSION: Even if not all criteria were found automatically due to the limitation to a frontal video, additional features are already emerging to better assess the athlete, especially in the back-to-sport process. The next process will be the calibration of the videos (height and width of the first and last hurdle) and a rescaling as they get closer (captured by the fixed order of the jumps).
© Copyright 2025 10th International Congress on Science and Skiing, January 28 - February 1, 2025, Val di Fiemme, Italy. Alle Rechte vorbehalten.