Real-time Semantic Scene Completion Via Feature Aggregation and Conditioned Prediction.
Xiaokang ChenYajie XingGang ZengPublished in: CoRR (2023)
Keyphrases
- real time
- live video
- prediction accuracy
- video rate
- camera tracking
- video images
- prediction model
- semantic similarity
- video sequences
- scene understanding
- prediction error
- text localization and recognition
- image rendering
- densely sampled
- semantic context
- feature vectors
- control system
- semantic web
- d scene
- feature set
- multiple images
- semantic information
- complex scenes
- image features
- feature correspondences
- appearance information
- input image
- scene recognition
- high level
- vision system
- single image
- dynamic scenes
- image regions
- spatial relations