基於壓縮感知演算法與機器學習之語者辨識

Speaker Recognition based on Compressive Sensing and Machine Learning

薛名祐(Ming-Yu Hsueh) ； 白明憲(Mingsian Bai)

《中華民國振動與噪音工程學會論文集》 2019 (2019/06) Pp. 76-79

摘要

本文實現語者辨識Speaker Recognition，其方法主要利用到機器學習Machine Learning（ML）以及壓縮感知原理Compressive Sensing，機器學習包含支撐向量機Support Vector Machine（SVM）、高斯混合模型－通用背景模型Gaussian mixture model-universal background model（GMM-UBM），壓縮感知原理則有貪婪算法Orthogonal Matching Pursuit（OMP）。

關鍵字

Speaker Identity ； Support Vector Machine ； Gaussian mixture model-universal background model ； Orthogonal Matching Pursuit

並列摘要

This thesis implements Speaker Identity, the method mainly uses Machine Learning (ML) and Compressive Sensing. The machine Learning includes Support Vector Machine(SVM) and Gaussian mixture model-universal background model (GMM-UBM). The principle of compressive sensing is greedy algorithm Orthogonal Matching Pursuit(OMP).

並列關鍵字

Speaker Identity ； Support Vector Machine ； Gaussian mixture model-universal background model ； Orthogonal Matching Pursuit