NeSF: Neural Semantic Fields for Generalizable Semantic Segmentation of 3D Scenes.
Suhani VoraNoha RadwanKlaus GreffHenning MeyerKyle GenovaMehdi S. M. SajjadiEtienne PotAndrea TagliasacchiDaniel DuckworthPublished in: Trans. Mach. Learn. Res. (2022)
Keyphrases
- d scene
- semantic segmentation
- single image
- conditional random fields
- depth map
- superpixels
- weakly supervised
- object categories
- scene classification
- optical flow
- object detection
- object classes
- object class
- pascal voc
- natural images
- training data
- image processing
- super resolution
- segmentation algorithm
- computer vision
- multiscale