Limited Training Data Robust Speech Recognition Using Kernel-Based Acoustic Models.
Martin SchaffönerSven E. KrügerEdin AndelicMarcel KatzAndreas WendemuthPublished in: ICASSP (1) (2006)
Keyphrases
- acoustic models
- speech recognition
- automatic speech recognition
- training data
- speech recognizer
- broadcast news
- hidden markov models
- noisy environments
- speaker independent
- discriminative training
- speech signal
- language model
- training process
- spoken language
- speaker identification
- support vector machine
- computer vision
- data sets
- kernel methods
- multi modal