Real-Time Semantic Scene Completion Via Feature Aggregation And Conditioned Prediction.
Xiaokang ChenYajie XingGang ZengPublished in: ICIP (2020)
Keyphrases
- real time
- video rate
- video images
- camera tracking
- prediction accuracy
- prediction model
- d scene
- natural language
- control system
- semantic web
- video sequences
- real environment
- semantic context
- densely sampled
- live video
- surveillance videos
- image features
- real scenes
- feature vectors
- high frame rate
- feature correspondences
- camera setup
- scene classification
- prediction error
- multiple images
- image sequences
- feature set
- multiple objects
- text localization and recognition
- motion features
- outdoor scenes
- dynamic scenes
- object recognition
- high level
- three dimensional
- feature selection