Semi-Supervised Enhancement and Suppression of Self-Produced Speech Using Correspondence between Air- and Body-Conducted Signals.
Moe TakadaShogo SekiPatrick Lumban TobingTomoki TodaPublished in: EUSIPCO (2020)
Keyphrases
- semi supervised
- audio signals
- speech recognition
- acoustic signals
- semi supervised learning
- signal processing
- fundamental frequency
- multi view
- supervised learning
- text to speech
- semi supervised classification
- noisy environments
- co training
- audio visual
- acoustic signal
- image enhancement
- human body
- image processing
- unlabeled data
- pairwise constraints
- labeled data
- cepstral features
- broadcast news
- automatic speech recognition
- body parts
- point correspondences
- unsupervised learning
- edge detection