Self Multi-head Attention For Speaker Recognition

Abstract

Most state-of-the-art Deep Learning (DL) approaches for speaker recognition work on a short utterance level. Given the speech signal, these algorithms extract a sequence of speaker embeddings from short segments and those are averaged to obtain an utterance level speaker representation. In this work we propose the use of an attention mechanism to obtain a discriminative speaker embedding given non fixed length speech utterances. Our system is based on a Convolutional Neural Network (CNN) that encodes short-term speaker features from the spectrogram and a self multi-head attention model that maps these representations into a long-term speaker embedding. The attention model that we propose produces multiple alignments from different subsegments of the CNN encoded states over the sequence. Hence this mechanism works as a pooling layer which decides the most discriminative features over the sequence to obtain an utterance level representation. We have tested this approach for the verificat

Stats

• citations69
• S2 citations—
• github stars0
• HF likes0
• heat score13.84
• arxiv keyindia2019self

Related papers

• Double Multi-head Attention For Speaker Verification (2020)8.09
• Self-attentive Multi-layer Aggregation With Feature Recalibration And Normalization For End-to-end Speaker Verification System (2020)0.00
• Attention Mechanism In Speaker Recognition: What Does It Learn In Deep Speaker Embedding? (2018)8.60
• Exploring A Unified Attention-based Pooling Framework For Speaker Verification (2018)6.77
• Saladnet: Self-attentive Multisource Localization In The Ambisonics Domain (2021)7.50
• End-to-end Attention Based Text-dependent Speaker Verification (2017)14.87
• Attention Back-end For Automatic Speaker Verification With Multiple Enrollment Utterances (2021)10.21
• Frequency And Temporal Convolutional Attention For Text-independent Speaker Recognition (2019)0.00