Exploring the Encoder Layers of Discriminative Autoencoders for LVCSR.
Pin-Tuan HuangHung-Shin LeeSyu-Siang WangKuan-Yu ChenYu TsaoHsin-Min WangPublished in: INTERSPEECH (2019)
Keyphrases
- deep belief networks
- unsupervised learning
- rate distortion
- denoising
- single layer
- bit rate
- multi layer
- low complexity
- video codec
- multiple layers
- video compression
- discriminative features
- spoken term detection
- feature selection
- hierarchical models
- feedforward neural networks
- coding efficiency
- image compression
- feature extraction