Graph Laplacian-Improved Convolutional Residual Autoencoder for Unsupervised Human Action and Emotion Recognition.
Giancarlo PaolettiCigdem BeyanAlessio Del BuePublished in: IEEE Access (2022)
Keyphrases
- emotion recognition
- human actions
- graph laplacian
- action recognition
- spatio temporal
- human motion
- human computer interaction
- audio visual
- activity recognition
- space time
- spectral clustering
- video sequences
- human activities
- facial expressions
- unsupervised learning
- sentiment analysis
- visual features
- bag of words
- multimedia
- visual data
- sparse coding
- search engine
- unlabeled data
- semi supervised learning
- supervised learning