Frequency-Warping Invariant Features for Automatic Speech Recognition.
Alfred MertinsJan RademacherPublished in: ICASSP (5) (2006)
Keyphrases
- invariant features
- automatic speech recognition
- speech recognition
- speech signal
- hidden markov models
- conversational speech
- image matching
- word error rate
- keypoints
- broadcast news
- speech retrieval
- shape features
- recognition errors
- texture images
- speech corpus
- spontaneous speech
- word recognition
- spoken words
- speaker adaptation
- color invariant
- viewpoint
- surface geometry
- pattern recognition