IDEAL
2003
Springer
14 years 22 days ago
2003
Springer
To reduce the high dimensionality required for training of feature vectors in speaker identification, we propose an efficient GMM based on local PCA with Fuzzy clustering. The prop...