Multi-View Features and Hybrid Reward Strategies for Vatex Video Captioning Challenge 2019.
Xinxin ZhuLongteng GuoPeng YaoJing LiuShichen LuZheng YuWei LiuHanqing LuPublished in: CoRR (2019)
Keyphrases
- multi view
- dynamic scenes
- single view
- multiple views
- depth video
- free viewpoint
- multiple cameras
- camera calibration
- d objects
- three dimensional
- depth map
- multi view clustering
- video sequences
- semi supervised
- stereo video
- multi view images
- feature extraction
- co training
- multi views
- multi view reconstruction
- multiple viewpoints
- multi view stereo
- view synthesis
- reinforcement learning
- feature space
- key frames
- scene reconstruction
- depth image based rendering
- surface reconstruction
- machine learning
- multi view video coding
- viewpoint
- feature vectors