Compact Speaker Embedding: lrx-Vector

Georges, Munir; Huang, Jonathan

doi:10.21437/interspeech.2020-2106

Cited by 11 publications

(8 citation statements)

References 27 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…edge or mobile devices, have not been fully studied. The topic was just recently investigated in [107,102,112].…”

Section: Discussion To the Networkmentioning

confidence: 99%

See 1 more Smart Citation

Speaker Recognition Based on Deep Learning: An Overview

Bai

Zhang

2020

Preprint

View full text Add to dashboard Cite

Speaker recognition is a task of identifying persons from their voices. Recently, deep learning has dramatically revolutionized speaker recognition. However, there is lack of comprehensive reviews on the exciting progress. In this paper, we review several major subtasks of speaker recognition, including speaker verification, identification, diarization, and robust speaker recognition, with a focus on deep-learning-based methods. Because the major advantage of deep learning over conventional methods is its representation ability, which is able to produce highly abstract embedding features from utterances, we first pay close attention to deep-learning-based speaker feature extraction, including the inputs, network structures, temporal pooling strategies, and objective functions respectively, which are the fundamental components of many speaker recognition subtasks. Then, we make an overview of speaker diarization, with an emphasis of recent supervised, end-to-end, and online diarization. Finally, we survey robust speaker recognition from the perspectives of domain adaptation and speech enhancement, which are two major approaches of dealing with domain mismatch and noise problems. Popular and recently released corpora are listed at the end of the paper.

show abstract

“…edge or mobile devices, have not been fully studied. The topic was just recently investigated in [107,102,112].…”

Section: Discussion To the Networkmentioning

confidence: 99%

“…The application of F-TDNN to deep embedding was also investigated [89,98,100]. Some other parameter reduction works can be found in [101,102]. Recently, [90] integrated TDNN with statistics pooling at each layer for compensating the variation of temporal context in the frame-level transforms.…”

Section: Deep Embedding: Network Structures and Inputsmentioning

confidence: 99%

Speaker Recognition Based on Deep Learning: An Overview

Bai

Zhang

2020

Preprint

View full text Add to dashboard Cite

show abstract

“…Mingote et al [40] explored the KD approach and presented a data augmentation technique to improve robustness of TD-SV systems. Georges et al [24] proposed a low-rank factorized version of the x-vector embedding network [5]. To design the lightweight models, Nunes et al [41] proposed a portable model called additive margin MobileNet1D (AM-MobileNet1D) for speaker identification on mobile devices, which uses raw waveform of speeches as input.…”

Section: Lightweight Architectures For Ti-svmentioning

confidence: 99%

“…On the other hand, with the application for access control in mobile devices [24], the design of SV systems tends to be lightweight and efficient. Existing DNN-based SV models, comprising of millions of parameters, require immense computational resources and are hard to achieve the lightweight goal.…”

Section: Introductionmentioning

confidence: 99%

RSKNet-MTSP: Effective and Portable Deep Architecture for Speaker Verification

Wu¹,

Guo²,

Zhao³

et al. 2021

Preprint

View full text Add to dashboard Cite

show abstract

“…On the other hand, with the widespread use of mobile devices [ 24 , 25 , 26 , 27 ], the design of speaker recognition systems tends to be light and efficient. However, existing models cannot be lighter, and the performance decreases drastically when made lighter.…”

Section: Introductionmentioning

confidence: 99%

ResSKNet-SSDP: Effective and Light End-To-End Architecture for Speaker Recognition

Deng

Jiang

et al. 2023

Sensors

View full text Add to dashboard Cite

In speaker recognition tasks, convolutional neural network (CNN)-based approaches have shown significant success. Modeling the long-term contexts and efficiently aggregating the information are two challenges in speaker recognition, and they have a critical impact on system performance. Previous research has addressed these issues by introducing deeper, wider, and more complex network architectures and aggregation methods. However, it is difficult to significantly improve the performance with these approaches because they also have trouble fully utilizing global information, channel information, and time-frequency information. To address the above issues, we propose a lighter and more efficient CNN-based end-to-end speaker recognition architecture, ResSKNet-SSDP. ResSKNet-SSDP consists of a residual selective kernel network (ResSKNet) and self-attentive standard deviation pooling (SSDP). ResSKNet can capture long-term contexts, neighboring information, and global information, thus extracting a more informative frame-level. SSDP can capture short- and long-term changes in frame-level features, aggregating the variable-length frame-level features into fixed-length, more distinctive utterance-level features. Extensive comparison experiments were performed on two popular public speaker recognition datasets, Voxceleb and CN-Celeb, with current state-of-the-art speaker recognition systems and achieved the lowest EER/DCF of 2.33%/0.2298, 2.44%/0.2559, 4.10%/0.3502, and 12.28%/0.5051. Compared with the lightest x-vector, our designed ResSKNet-SSDP has 3.1 M fewer parameters and 31.6 ms less inference time, but 35.1% better performance. The results show that ResSKNet-SSDP significantly outperforms the current state-of-the-art speaker recognition architectures on all test sets and is an end-to-end architecture with fewer parameters and higher efficiency for applications in realistic situations. The ablation experiments further show that our proposed approaches also provide significant improvements over previous methods.

show abstract

Compact Speaker Embedding: lrx-Vector

Cited by 11 publications

References 27 publications

Speaker Recognition Based on Deep Learning: An Overview

Speaker Recognition Based on Deep Learning: An Overview

RSKNet-MTSP: Effective and Portable Deep Architecture for Speaker Verification

ResSKNet-SSDP: Effective and Light End-To-End Architecture for Speaker Recognition

Contact Info

Product

Resources

About