Trapezoid-structured LSTM with segregated gates and bridge joints for video frame inpainting.
Ting-Hui ChiangYun-Tang LinJaden Chao-Ho LinYu-Chee TsengPublished in: Vis. Comput. (2024)
Keyphrases
- video frames
- video sequences
- body parts
- spatial and temporal
- video data
- video segmentation
- inter frame
- video streams
- video content
- foreground objects
- human body
- saliency map
- human motion
- consecutive frames
- video images
- texture synthesis
- image restoration
- neural network
- degrees of freedom
- image inpainting
- video shots
- video clips
- recurrent neural networks
- key frames
- denoising
- static images
- computer vision
- image sequences
- three dimensional
- image segmentation
- video dataset