Automatic assignment of anchoring points on vowel templates for defining correspondence between time-frequency representations of speech samples.
Toru TakahashiMasashi NishiToshio IrinoHideki KawaharaPublished in: INTERSPEECH (2006)
Keyphrases
- speech signal
- sample points
- speech synthesis
- vocal tract
- speech recognition
- formant frequencies
- prosodic features
- semi automatic
- speech recognition systems
- point sets
- data points
- fully automatic
- point correspondences
- fundamental frequency
- speaker independent
- image matching
- training samples
- text to speech
- speech recognizer
- spoken language
- data sets
- automatic speech recognition
- audio visual
- convex hull
- speaker recognition
- point pairs
- broadcast news
- digit recognition
- epipolar lines
- linear prediction
- point to point correspondences
- speech music discrimination