From CAD models to soft point cloud labels: An automatic annotation pipeline for cheaply supervised 3D semantic segmentation.
Galadrielle Humblot-RenauxSimon Buus JensenAndreas MøgelmosePublished in: CoRR (2023)
Keyphrases
- semantic segmentation
- automatic annotation
- point cloud
- cad model
- weakly supervised
- image annotation
- superpixels
- object categories
- structure from motion
- point sets
- image understanding
- low level features
- supervised learning
- learning algorithm
- semi supervised
- feature selection
- object classes
- multiscale
- image sequences
- wordnet
- graph cuts
- object detection
- d objects
- pairwise
- feature space
- computer vision