Taketomi, TakafumiYoshitake, YasuhideYamamoto, GoshiroSandor, ChristianKato, HirokazuTony Huang and Arindam Dey2017-11-212017-11-212017978-3-03868-052-9https://doi.org/10.2312/egve.20171376https://diglib.eg.org:443/handle/10.2312/egve20171376We propose a method for visualizing 3D ground reaction forces for sprint training. Currently, sprinters can check their 3D ground force data using a 2D graph representation. In order to check the relationship between 3D ground force and their sprint form, they must check the 2D graph and a training video repeatedly. To allow simultaneous observation of the 2D graph and the training video, we use a mixed reality technology to overlay 3D ground reaction force onto the training video. In this study, we focus on 2D-3D registration between the image sequence and 3D ground reaction data. We achieved 2D-3D registration by using a constrained bundle adjustment approach. In the experiment, we apply our method to the training videos. The results confirm that our method can correctly overly 3D ground reaction force onto the videos.Computing methodologiesMixed / augmented realityHumancentered computingVisualization10.2312/egve.2017137625-26