Variable-rate Neural Speech Compression with Multi-scale Feature Extraction and Improved Entropy Modeling.
Shaohan SunYuzhuo KongTong ChenZhan MaPublished in: DCC (2024)
Keyphrases
- texture features
- feature extraction
- local binary pattern
- multiscale
- texture analysis
- mathematical modeling
- data compression
- wavelet transform
- speech recognition
- natural images
- neural model
- compression algorithm
- pattern classification
- scale space
- image compression
- mutual information
- preprocessing
- image segmentation
- neural network
- keypoint detection
- speaker identification
- speech signal
- coarse to fine
- feature selection