3D reconstruction with auto-selected keyframes based on depth completion correction and pose fusion.
Fangzheng TianYongbin GaoZhijun FangJia GuShuqun YangPublished in: J. Vis. Commun. Image Represent. (2021)
Keyphrases
- key frames
- video retrieval
- pose recovery
- pose estimation
- visual features
- visual content
- low level features
- d objects
- video content
- video data
- three dimensional
- shot detection
- visual cues
- depth information
- data fusion
- depth map
- video sequences
- video clips
- feature vectors
- video summarization
- d scene
- structure from motion
- semantic concepts
- news stories
- databases
- visual information
- low level
- high resolution
- depth estimation
- representative images