GPT4Motion: Scripting Physical Motions in Text-to-Video Generation via Blender-Oriented GPT Planning.
Jiaxi LvYi HuangMingfu YanJiancheng HuangJianzhuang LiuYifan LiuYafei WenXiaoxin ChenShifeng ChenPublished in: CoRR (2023)
Keyphrases
- object motion
- video sequences
- body motions
- motion capture data
- motion features
- image sequences
- space time
- human motion
- motion model
- dynamic scenes
- moving objects
- optical flow
- motion patterns
- text generation
- mechanical devices
- motion planning
- motion trajectories
- motion analysis
- input video
- motion estimation
- video data
- spatial and temporal
- motion sequences
- motion capture
- key frames
- image motion
- temporal filtering
- camera movement
- multibody
- visual data
- moving camera
- collision free
- camera motion
- video scene
- feature trajectories
- natural language descriptions
- spatio temporal
- articulated body
- motion field
- motion recognition
- video segments
- multimedia
- video frames
- activity recognition
- motion segmentation
- video retrieval
- temporal continuity
- human actions
- video surveillance
- layered representation
- rigid motion
- video objects
- video search
- human walking
- visual cues
- video streams
- dynamic textures
- computer vision
- video shots
- wide area motion imagery
- multi view
- human body
- structure from motion
- motion boundaries
- body movements
- humanoid robot
- video analysis
- motion primitives
- path planning
- background subtraction
- human activities
- video clips
- point trajectories
- motion parameters