Unsupervised Learning of Video Representations via Dense Trajectory Clustering.
Pavel TokmakovMartial HebertCordelia SchmidPublished in: CoRR (2020)
Keyphrases
- unsupervised learning
- data clustering
- supervised learning
- unsupervised manner
- cluster analysis
- video data
- semi supervised
- deep belief networks
- video sequences
- real time
- video content
- self organizing maps
- clustering algorithm
- dimensionality reduction
- point trajectories
- k means
- feature selection
- text classification
- minimum message length
- space time
- video streams
- data sets
- data representations
- video scene
- multimedia
- deep learning
- video analysis
- video clips
- human activities
- pattern recognition
- expectation maximization
- decision trees
- image sequences
- video shots
- training data
- video frames
- bayesian networks
- high frame rate
- data points