Spatio-temporal categorization for first-person-view videos using a convolutional variational autoencoder and Gaussian processes.
Masatoshi NaganoTomoaki NakamuraTakayuki NagaiDaichi MochihashiIchiro KobayashiPublished in: Frontiers Robotics AI (2022)
Keyphrases
- gaussian processes
- spatio temporal
- gaussian process
- gaussian process regression
- human actions
- covariance function
- text categorization
- video sequences
- moving objects
- gaussian process models
- restricted boltzmann machine
- multi task
- image segmentation
- spatial and temporal
- human pose estimation
- image sequences
- hyperparameters
- deep learning
- latent variables
- video data
- model selection
- semi supervised
- active learning