Hybridized estimations of support vector machine free parameters C and γ using a fuzzy learning strategy for microphone array-based speaker recognition in a Kinect sensor-deployed environment.
Ing-Jr DingJia-Yi ShiPublished in: Multim. Tools Appl. (2017)
Keyphrases
- learning strategies
- free parameters
- speaker recognition
- support vector machine
- kinect sensor
- gaussian mixture model
- probabilistic neural network
- online learning
- active learning
- speaker verification
- vector quantization
- depth information
- machine learning
- speaker identification
- multimedia
- noisy environments
- speech recognition
- depth images
- speech signal
- mixture model
- training data
- feature extraction
- feature selection