Unified ASR system using LGM-based source separation, noise-robust feature extraction, and word hypothesis selection.
Yusuke FujitaRyoichi TakashimaTakeshi HommaRintaro IkeshitaYohei KawaguchiTakashi SumiyoshiTakashi EndoMasahito TogamiPublished in: ASRU (2015)
Keyphrases
- source separation
- feature extraction
- noisy environments
- blind source separation
- independent components analysis
- noise level
- single channel
- face recognition
- dimensionality reduction
- image classification
- speech recognition
- independent component analysis
- automatic speech recognition
- feature vectors
- multiscale
- model selection
- frequency domain
- noise reduction
- principal component analysis
- wavelet transform
- additive noise