FutureGAN: Anticipating the Future Frames of Video Sequences using Spatio-Temporal 3d Convolutions in Progressively Growing Autoencoder GANs.
Sandra AignerMarco KörnerPublished in: CoRR (2018)
Keyphrases
- video sequences
- spatio temporal
- moving objects
- video frames
- frame rate
- image sequences
- space time
- spatial and temporal
- temporal domain
- human motion
- motion trajectories
- human actions
- key frames
- consecutive frames
- video objects
- motion estimation
- spatial temporal
- object tracking
- long term
- dynamic scenes
- video segmentation
- three dimensional
- spatio temporal databases
- foreground objects
- video content
- fourier transform
- camera motion
- motion vectors
- video segments
- spatio temporal data
- real world
- motion segmentation
- temporal information
- background subtraction
- action recognition
- reference frame
- motion patterns
- multi view
- single frame
- high speed
- temporal coherence
- optical flow
- neural network
- adjacent frames
- interesting events
- objects in video sequences