Mapping Sparse Representation to State Likelihoods in Noise-Robust Automatic Speech Recognition.
Katariina MahkonenAntti HurmalainenTuomas VirtanenJort F. GemmekePublished in: INTERSPEECH (2011)
Keyphrases
- sparse representation
- automatic speech recognition
- noisy environments
- dictionary learning
- norm minimization
- speech recognition
- sparse coding
- face recognition
- image classification
- compressed sensing
- compressive sensing
- image representation
- speech signal
- image patches
- hidden markov models
- test images
- noise reduction
- conversational speech
- word error rate
- sparsity constraints
- speech retrieval
- random projections
- signal to noise ratio
- high dimensional data
- broadcast news
- maximum likelihood
- sparse coefficients
- machine learning
- signal processing
- information retrieval