Latent discriminative representation learning for speaker recognition

• Published in: Frontiers of information technology & electronic engineering Vol. 22; no. 5; pp. 697 - 708
• Main Authors: Huang, Duolin, Mao, Qirong, Ma, Zhongchen, Zheng, Zhishen, Routryar, Sidheswar, Ocquaye, Elias-Nii-Noi
• Format: Journal Article
• Language: English
• Published: Hangzhou Zhejiang University Press 01.05.2021; Springer Nature B.V; Jiangsu Key Laboratory of Security Technology for Industrial Cyberspace,Zhenjiang 212013,China; School of Computer Science and Communication Engineering,Jiangsu University,Zhenjiang 212013,China%School of Computer Science and Communication Engineering,Jiangsu University,Zhenjiang 212013,China
• Subjects: Communications Engineering; Computer Hardware; Computer Science; Computer Systems Organization and Communication Networks; Datasets; Deep learning; Dictionaries; Discriminant analysis; Electrical Engineering; Electronics and Microelectronics; Instrumentation; Learning; Lookup tables; Mathematical analysis; Networks; Neural networks; Representations; Speech; Speech recognition; Teaching methods; Voice recognition; Speaker embedding lookup table; 潜在可区分性表征学习; Linear mapping matrix; 说话人识别; Latent discriminative representation learning; Speaker recognition; TP391.4; 线性映射矩阵; 说话人嵌入查找表
• ISSN: 2095-9184; 2095-9230
• DOI: 10.1631/FITEE.1900690

Extracting discriminative speaker-specific representations from speech signals and transforming them into fixed length vectors are key steps in speaker identification and verification systems. In this study, we propose a latent discriminative representation learning method for speaker recognition. We mean that the learned representations in this study are not only discriminative but also relevant. Specifically, we introduce an additional speaker embedded lookup table to explore the relevance between different utterances from the same speaker. Moreover, a reconstruction constraint intended to learn a linear mapping matrix is introduced to make representation discriminative. Experimental results demonstrate that the proposed method outperforms state-of-the-art methods based on the Apollo dataset used in the Fearless Steps Challenge in INTERSPEECH2019 and the TIMIT dataset.