From Real to Synthetic and Back: Synthesizing Training Data for Multi-Person Scene Understanding.
Igor KviatkovskyNadav BhonkerGérard G. MedioniPublished in: CoRR (2020)
Keyphrases
- scene understanding
- training data
- object detection
- object recognition
- vision system
- d scene
- robot navigation
- scene recognition
- video surveillance
- scene labeling
- geometric reasoning
- scene interpretation
- image features
- decision trees
- image parsing
- learning algorithm
- data sets
- maximum likelihood
- labeled data
- depth map
- feature vectors
- viewpoint
- training set
- support vector
- computer vision