computer-vision-18
3D Semantic Scene Completion
3260 papers • 126 benchmarks • 313 datasets
This task was introduced in "Semantic Scene Completion from a Single Depth Image" (https://arxiv.org/abs/1611.08974) at CVPR 2017 . The target is to infer the dense 3D voxelized semantic scene from an incompleted 3D input (e.g. point cloud, depth map) and an optional RGB image. A recent summary can be found in the paper "3D Semantic Scene Completion: a Survey" (https://arxiv.org/abs/2103.07466), published at IJCV 2021.
(Image credit: Open Source)
Benchmarks
These leaderboards are used to track progress in 3d-semantic-scene-completion-18
Trend
Dataset
Best Model
Actions
NYUv2
NYUv2
SemanticKITTI
SemanticKITTI
KITTI-360
KITTI-360
Libraries
iUse these libraries to find 3d-semantic-scene-completion-18 models and implementations
Datasets
No datasets available.
