| CPC G10L 17/02 (2013.01) [G10L 15/08 (2013.01); G10L 25/51 (2013.01)] | 12 Claims |
|
1. A voice recognition method, comprising:
collecting a plurality of voice signals;
extracting voiceprint features of each of the voice signals;
performing a data process on the voiceprint features, to convert the voiceprint features into a N-dimensional matrix, and N is an integer greater than or equal to 2;
performing a feature normalization process on the N-dimensional matrix to scale the voiceprint features proportionally to fall within a specific interval to obtain a plurality of voiceprint data;
classifying the voiceprint data to generate a clustering result; and
finding out a centroid of each cluster according to the clustering result, and registering the voiceprint data adjacent to each of the centroid;
wherein the step of classifying the voiceprint data further comprises:
calculating a number of the clusters and adjacent slopes by an elbow method according to the voiceprint data;
generating the clustering result of classification by a hierarchical clustering algorithm using the calculated number of clusters when the slope changes suddenly; and
generating the clustering result of classification directly by the hierarchical clustering algorithm without using the calculated number of clusters when the slope does not change suddenly.
|