MotionBooth: Motion-Aware Customized Text-to-Video Generation.
Jianzong WuXiangtai LiYanhong ZengJiangning ZhangQianyu ZhouYining LiYunhai TongKai ChenPublished in: CoRR (2024)
Keyphrases
- space time
- temporal filtering
- key frames
- object motion
- input video
- text generation
- motion features
- spatial and temporal
- video sequences
- motion analysis
- moving camera
- video data
- temporal continuity
- low frame rate
- image sequences
- static images
- news video
- dynamic textures
- information retrieval
- video scene
- video representation
- natural language descriptions
- video segments
- video search
- motion estimation
- shot change detection
- temporal consistency
- text detection
- visual cues
- multimedia
- visual data
- camera motion
- video frames
- video content
- video analysis
- keywords
- video streams
- temporal coherence
- motion trajectories
- surveillance videos
- motion model
- successive frames
- video objects
- video retrieval
- human motion
- spatio temporal
- video footage
- video signals
- motion detection
- motion patterns
- video database
- multimedia documents
- image frames
- camera movement
- periodic motion
- dynamic scenes
- layered representation
- visual information
- motion capture data
- video clips
- video shots
- temporal information
- motion capture
- text mining
- multiview video
- event detection
- optical flow
- single frame
- motion field
- reference frame