Current motion forecasting approaches typically train a deep end-to-end model from the source domain data, and then apply it directly to target subjects. Despite promising results, they remain ...
Some results have been hidden because they may be inaccessible to you
Show inaccessible results