Investigating accuracy of pitch-accent annotations in neural network-based speech synthesis and denoising effects.
Hieu-Thi LuongXin WangJunichi YamagishiNobuyuki NishizawaPublished in: CoRR (2018)
Keyphrases
- speech synthesis
- denoising
- speech recognition
- image denoising
- high accuracy
- neural network
- total variation
- noisy images
- text to speech
- computational cost
- semantic annotation
- vocal tract
- error rate
- prediction accuracy
- pattern recognition
- wavelet domain
- object recognition
- prosodic features
- inter annotator agreement
- natural images
- hidden markov models
- multiscale
- image processing