GIO: A Timbre-informed Approach for Pitch Tracking in Highly Noisy Environments.
Xiaoheng SunXia LiangQiqi HeBilei ZhuZejun MaPublished in: ICMR (2022)
Keyphrases
- noisy environments
- acoustic features
- speaker verification
- speech signal
- speech recognition
- noise reduction
- speech enhancement
- automatic speech recognition
- digit recognition
- background noise
- particle filter
- speaker identification
- significantly outperformed
- mean shift
- maximum likelihood
- wavelet transform
- computer vision
- feature space
- image segmentation
- neural network