OpenXtract: A Blender Add-On for the Accelerated Extraction of the Objects of Interest

In this contribution, we extended and evaluated five open source point cloud partition algorithms. The extensions leverage the edges of a mesh in order to approximate geodesic distances. They achieved higher accuracies and partition sizes as their point cloud counterpart due to more precise eigenfeatures. The usage of the extensions is recommended in order to get sharper edges. The partition algorithms are made open source and can be used in Python and C++. In addition, we propose an open source Blender add-on, called OpenXtract, where the partition algorithms and a PointNet++ semantic segmentation can be applied. We validated the usability of OpenXtract by interviewing three 3D experts. According to the interviews, OpenXtract is extendable and can be integrated into 3D design workflows.