Speech Reconstruction from MFCC Based on Nonnegative and Sparse Priors.
Gang MinXiongwei ZhangJibin YangXia ZouZhi song PanPublished in: IEICE Trans. Fundam. Electron. Commun. Comput. Sci. (2015)
Keyphrases
- speech recognition
- speech signal
- compressed sensing
- compressive sensing
- speaker identification
- signal reconstruction
- image reconstruction
- speaker recognition
- speech recognition systems
- tensor factorization
- sparse representation
- sparse reconstruction
- automatic speech recognition
- reconstruction error
- linear programming
- binary matrices
- mel frequency cepstral coefficients
- hidden markov models
- three dimensional
- speech recognizer
- bayesian framework
- acoustic features
- speaker verification
- speaker diarization
- natural images
- least squares
- cepstral coefficients
- objective function
- noisy speech
- speech synthesis
- discrete tomography
- reconstruction method
- spoken language
- audio features
- pattern recognition
- high dimensional
- dictionary learning
- high resolution
- feature extraction
- broadcast news
- feature vectors
- speech enhancement
- language model
- gaussian mixture model
- sparsity constraints
- maximum a posteriori
- prior information
- sparse coding
- audio visual
- random projections
- spectral features
- noisy environments