CoordViT: A Novel Method of Improve Vision Transformer-Based Speech Emotion Recognition using Coordinate Information Concatenate.
Jeong-Yoon KimSeung-Ho LeePublished in: ICEIC (2023)
Keyphrases
- prior knowledge
- high accuracy
- segmentation method
- clustering method
- cost function
- neural network
- prior information
- high precision
- error rate
- detection method
- experimental evaluation
- dynamic programming
- preprocessing
- image processing
- information extraction
- domain knowledge
- human computer interaction
- significant improvement
- similarity measure
- computer vision
- real time
- statistical information