Speech Enhancement in Quasi-Periodic Noises Using Improved Spectral Subtraction Based on Adaptive Sampling.
Elias AzarovMaxim VashkevichAlexander A. PetrovskyPublished in: SPECOM (2015)
Keyphrases
- speech enhancement
- spectral subtraction
- adaptive sampling
- quasi periodic
- speech signal
- speech recognition
- noise reduction
- noisy environments
- signal to noise ratio
- noisy speech
- single channel
- monte carlo
- linear prediction
- background noise
- vocal tract
- dynamic time warping
- automatic speech recognition
- wiener filter
- random sampling
- hidden markov models
- smoothing algorithm
- language model
- information retrieval
- principal component analysis
- noisy images
- image processing
- data mining