Unsupervised Single-Channel Singing Voice Separation with Weighted Robust Principal Component Analysis Based on Gammatone Auditory Filterbank and Vocal Activity Detection.
Feng LiYujun HuLingling WangPublished in: Sensors (2023)
Keyphrases
- single channel
- activity detection
- sound source
- filter bank
- robust principal component analysis
- speech signal
- audio features
- signal processing
- subband
- frequency domain
- wavelet decomposition
- multi channel
- multiresolution
- independent component analysis
- speech recognition
- multiscale
- semi supervised
- audio visual
- machine learning
- wavelet transform
- bit rate
- spatial domain
- wavelet domain
- noise reduction
- hidden markov models
- wavelet coefficients
- image compression