Leveraging Semantic Scene Characteristics and Multi-Stream Convolutional Architectures in a Contextual Approach for Video-Based Visual Emotion Recognition in the Wild.
Ioannis PikoulisPanagiotis Paraskevas FilntisisPetros MaragosPublished in: FG (2021)
Keyphrases
- multi stream
- emotion recognition
- audio visual
- visual data
- visual information
- contextual information
- emotional speech
- hidden markov models
- semantic information
- video sequences
- image sequences
- visual features
- facial expressions
- input image
- sentiment analysis
- human computer interaction
- low level
- emotional state
- video data
- facial images
- high dimensional
- natural language
- image classification
- computer vision