Abstract: Supervised methods on 3D medical image segmentation need large amounts of annotated data, but annotating is time-consuming. Also, existing 3D segmentation methods capture more global ...
Abstract: We present a comprehensive framework for egocentric interaction recognition using markerless 3D annotations of two hands manipulating objects. To this end, we propose a method to create a ...
Some results have been hidden because they may be inaccessible to you
Show inaccessible results
Feedback