VideoDirectorGPT: Consistent Multi-scene Video Generation via LLM-Guided Planning.
Han LinAbhay ZalaJaemin ChoMohit BansalPublished in: CoRR (2023)
Keyphrases
- video sequences
- video images
- video scene
- d scene
- dynamic scenes
- video data
- moving camera
- scene change detection
- visual data
- multiple frames
- single frame
- video streams
- object detection and tracking
- video footage
- multiple video streams
- object motion
- input video
- three dimensional
- image mosaics
- motion features
- surveillance videos
- moving objects
- multiple images
- multimedia
- high resolution video
- single image
- live video
- video frames
- image frames
- video analysis
- real scenes
- video camera
- image sequences
- video clips
- space time
- video retrieval
- video content
- scene analysis
- video surveillance
- scene understanding
- dynamic textures
- multiple objects
- neighboring frames
- stationary camera
- planning problems
- video shots
- human actions
- object tracking
- heuristic search
- crowded scenes
- complex scenes
- layered representation
- combining information from multiple
- human activities
- static images