BAE-Net: A Low complexity and high fidelity Bandwidth-Adaptive neural network for speech super-resolution.
Guochen YuXiguang ZhengNan LiRunqiang HanChengshi ZhengChen ZhangChao ZhouQi HuangBing YuPublished in: CoRR (2023)
Keyphrases
- low complexity
- high fidelity
- super resolution
- high resolution
- motion estimation
- video streaming
- low resolution
- high quality
- image reconstruction
- image restoration
- low resolution images
- real time
- multi frame
- computational complexity
- audio visual
- video conferencing
- video compression
- image sequences
- video coding
- motion compensation
- distributed video coding
- spatial resolution
- motion vectors
- depth map
- video sequences
- low spatial resolution
- motion compensated
- multiple description coding
- high frequency
- inter frame
- computer vision
- machine learning
- rate distortion
- three dimensional
- spatial domain
- motion model
- optical flow
- spatio temporal
- compressed video