어렵긴 하지만 가능할 것 같다. 'A robust 3D-2D interactive tool for scene segmentation' 논문을 살펴보면 좋을 것 같다. 비슷한 task 는 semantic kitty, 3d instance segmentation 이 있다. -주영준

Real time stereo 3D semantic mapping in large scale environments. Code: https://github.com/shichaoy/semantic_3d_mappingPaper: http://www.frc.ri.cmu.edu/~syan...

The current state-of-the-art on SemanticKITTI is AF2S3Net. See a full comparison of 26 papers with code.

Image: [OccuSeg](https://arxiv.org/pdf/2003.06537v3.pdf)

Recent advances of 3D acquisition devices have enabled large-scale acquisition of 3D scene data. Such data, if completely and well annotated, can serve as useful ingredients for a wide spectrum of computer vision and graphics works such as data-driven modeling and scene understanding, object detection and recognition.