Latent-Space Disentanglement with Untrained Generator Networks for the Isolation of Different Motion Types in Video Data.
Abdullah AbdullahMartin HollerKarl KunischMalena Sabate LandmanPublished in: SSVM (2023)
Keyphrases
- video data
- latent space
- key frames
- video sequences
- gaussian process latent variable models
- multimedia
- video camera
- motion estimation
- human motion
- motion analysis
- image sequences
- space time
- video frames
- moving objects
- latent variables
- dynamic scenes
- video shots
- camera motion
- low dimensional
- high dimensional
- motion model
- object recognition
- three dimensional
- gaussian mixture model
- feature extraction
- object motion
- lower dimensional
- gaussian process
- optical flow
- matrix factorization
- co occurrence
- manifold learning
- generative model
- human body
- missing data