An Improved Combined-Dictionary Method for the Judgment of Voiced and Unvoiced Sounds
2020 International Conference on Mathematics and Computers in Science and Engineering (MACISE)(2020)
摘要
To solve the problem of the confusion in atoms between voiced and unvoiced dictionaries, an improved Combined-Dictionary method of the judgment in voiced and unvoiced sounds is proposed. Combining the correlation coefficient method and Mahalanobis distance, a new approach is explored to analyze the similarity of atoms. By analyzing the correlation in corresponding atoms, the appropriate threshold value is selected and the clean combined-dictionaries are obtained. The results have indicted, under the same conditions, the improved combined-dictionary method has higher accuracy and faster computing speed than the original method.
1
更多查看译文
关键词
Mahalanobis distance,correlation coefficient method,sparse representation,dictionary learning
AI 理解论文
溯源树
样例
生成溯源树,研究论文发展脉络
Chat Paper
正在生成论文摘要