VORNet: Spatio-Temporally Consistent Video Inpainting for Object Removal.
Ya-Liang ChangZhe Yu LiuWinston H. HsuPublished in: CVPR Workshops (2019)
Keyphrases
- temporally consistent
- temporal consistency
- optical flow
- superpixels
- depth map
- video streams
- motion parameters
- key frames
- multiple objects
- object segmentation
- d objects
- video data
- multimedia
- multiscale
- image sequences
- bounding box
- high quality
- depth information
- input image
- temporal constraints
- target object
- video sequences
- video frames
- visual features
- multi view
- higher order
- object classes
- motion estimation
- long range
- ground truth
- moving objects
- optical flow estimation