Mixtures of local dictionaries for unsupervised speech enhancement

Keyphrases

• Unsupervised Speech Enhancement 100%
• Local Dictionary 100%
• Non-negative Matrix Factorization 75%
• Sparsity Constraint 25%
• Dictionary 25%
• Local Structure 25%
• Noise Type 25%
• State-of-the-art Technology 25%
• Single-channel Speech Enhancement 25%
• Number of Blocks 25%
• Speaker Identity 25%
• Block Sparsity 25%
• Locality Preservation 25%
• Signal-to-Distortion Ratio 25%

Computer Science

• Speech Enhancement 100%
• nonnegative matrix factorization 100%
• Sparsity 33%
• Local Structure 33%
• Single Channel Speech Enhancement 33%