An expectation-maximization eigenvector clustering approach to direction of arrival estimation of multiple speech sources.
Xiong XiaoShengkui ZhaoThi Ngoc Tho NguyenDouglas L. JonesEng Siong ChngHaizhou LiPublished in: ICASSP (2016)
Keyphrases
- direction of arrival
- doa estimation
- expectation maximization
- signal subspace
- em algorithm
- k means
- clustering algorithm
- data clustering
- clustering method
- speech recognition
- parameter estimation
- canonical correlation analysis
- unsupervised learning
- spectral clustering
- least squares
- cluster analysis
- maximum likelihood
- infrared
- covariance matrix
- spectral methods
- data points