Journal of Jilin University(Engineering and Technology Edition) ›› 2024, Vol. 54 ›› Issue (11): 3318-3326.doi: 10.13229/j.cnki.jdxbgxb.20221635

Previous Articles    

Voiceprint recognition method based on novel loss function DV-Softmax

Yi CAO(),Ping LI,Wei-guan WU,Yu XIA,Qing-yuan GAO   

  1. School of Mechanical Engineering，Jiangnan University，Wuxi 214122，China
  • Received:2022-12-28 Online:2024-11-01 Published:2025-04-24

RICH HTML

 2   

PDF (PC)

137

Abstract:

In view of the problems that the classification loss function of existing models in the field of voiceprint recognition cannot effectively distinguish the separability between categories and lack of attention to the quality of voiceprint data， a new classification loss function DV-Softmax is proposed in this paper. Firstly， the working principle of the existing boundary loss function in voiceprint field is introduced. Secondly， the mining loss function in the field of object detection is introduced， and the concept of fuzzy sample is proposed based on it. Then， the MV-Softmax loss function is introduced in the field of face recognition， and fuzzy samples are added to make it adaptive to emphasize the difference between different samples and guide the feature learning. Finally， the voicing recognition was studied on Voxceleb1 and SITW data respectively. The experimental results show that compared with the existing boundary loss function， the equal error rate of DV-Softmax is reduced by 8% and 5.4%， respectively， which verifies that the DV-Softmax loss function effectively solves the separability between categories and concerns the quality of sample voice print data， and has a good performance in the field of voice print recognition。

Key words: deep learning, voiceprint recognition, loss function, information mining

CLC Number: 

  • TN912.34

Fig.1

Schematic diagram of weight indicator function"

Fig.2

The framework of the proposed DV-Softmax"

Fig.3

Geometric interpretation from featureperspective"

Table 1

Pseudocode table of DV-Softmax"

Algorithm：DV-Softmax

Input：Training set x with its annotated label y.

Epochs， Adam， The hyper-parameter t.

Initialization： Randomly initialize the parameter Θ in convolution layers and W in the last fully connected layer.

function GO （ ）：

for epoch in Epochs to do

Training model use Adam
Forward： According to the indication of different examples Eq. （11）， we compute the DV-Softmax loss by Eq. （12）；
Backward： Update the parameters W and Θ by Adam
end for
returnW and Θ
end function
Output： Parameters Θ and W.

Table 2

Performance comparison of different loss functions on Voxceleb1 test sets"

损失函数EER/%minDCF（P=0.1）minDCF（P=0.01）minDCF（P=0.001）
Softmax3.890.2670.4360.483
A-Softmax3.020.1930.3520.405
AM-Softmax2.720.1500.3080.350
AAM-Softmax2.490.1320.2690.312
F-Softmax3.930.2430.4180.463
MV-AAM-Softmax-f2.380.1260.2580.298
MV-AAM-Softmax-a2.340.1210.2540.288
D-AAM-Softmax2.440.1230.2560.295
D-F-Softmax3.710.2180.4100.443
DV-AAM-Softmax-f2.320.1180.2420.278
DV-AAM-Softmax-a2.290.1130.2380.272

Fig.4

Comparison of DET curves of different loss functions on Voxceleb1 data set"

Table 3

Performance comparison of different loss functions on SITW test sets"

损失函数EER/%minDCF（P=0.1）minDCF（P=0.01）minDCF（P=0.001）
Softmax6.290.2960.5460.752
A-Softmax5.210.2560.5100.706
AM-Softmax4.730.2260.4330.644
AAM-Softmax3.910.1790.3480.551
F-Softmax6.350.3120.5530.763
MV-AAM-Softmax-f3.780.1680.3340.542
MV-AAM-Softmax-a3.750.1650.3230.538
D-AAM-Softmax3.870.1740.3430.535
D-F-Softmax6.240.2930.5430.727
DV-AAM-Softmax-f3.730.1650.3200.521
DV-AAM-Softmax-a3.700.1620.3160.515

Fig.5

Comparison of DET curves of different loss functions on SITW data set"

1 Ranjan R, Castillo C D, Chellappa R. L2-constrained softmax loss for discriminative face verification[J]. Arxiv Preprint, 2017, 3: No.170309507.
2 Liu W, Wen Y, Yu Z, et al. Sphereface: deep hypersphere embedding for face recognition[C]∥Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, Hololulu, USA, 2017: 212-220.
3 Wang F, Cheng J, Liu W, et al. Additive margin softmax for face verification[J]. IEEE Signal Processing Letters, 2018, 25(7): 926-930.
4 Deng J, Guo J, Xue N, et al. Arcface: additive angular margin loss for deep face recognition[C]∥Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, Long Beach, USA, 2019: 4690-4699.
5 Thienpondt J, Desplanques B, Demuynck K. Cross-lingual speaker verification with domain-balanced hard prototype mining and language-dependent score normalization[J]. Arxiv Preprint, 2020, 7: No. 200707689.
6 Li X, Wang W, Wu L J, et al. Generalized focal loss: learning qualified and distributed bounding boxes for dense object detection[J]. Advances in Neural Information Processing Systems, 2020, 33: 21002-21012.
7 Ma C, Sun H, Zhu J, et al. Normalized maximal margin loss for open-set image classification[J]. IEEE Access, 2021, 9: 54276-54285.
8 Lee J, Wang Y, Cho S. Angular margin-mining softmax loss for face recognition[J]. IEEE Access, 2022, 10: 43071-43080.
9 Boutros F, Damer N, Kirchbuchner F, et al. Elasticface: elastic margin loss for deep face recognition[C]∥ Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, New Orleans, USA, 2022: 1578-1587.
10 Wang X, Zhang S, Wang S, et al. Mis-classified vector guided softmax loss for face recognition[C]∥ Proceedings of the AAAI Conference on Artificial Intelligence, New York, USA, 2020, 34(7): 12241-12248.
11 Nagrani A, Chung J S, Zisserman A, et al. Voxceleb: a large-scale speaker identification dataset[J]. Arxiv Preprint, 2017, 6: No.170608612.
12 Mclaren M, Ferrer L, Castan D, et al. The speakers in the wild (SITW) speaker recognition database[C]∥ Proceedings of the Interspeech, San Francisco, USA, 2016: 818-822.
13 Shrivastava A, Gupta A, Girshick R. Training region-based object detectors with online hard example mining[C]∥Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, Las Vegas, USA, 2016: 761-769.
14 Lin T Y, Goyal P, Girshick R, et al. Focal loss for dense object detection[C]∥Proceedings of the IEEE International Conference on Computer Vision, Venice, Italy, 2017: 2980-2988.
15 Desplanques B, Thienpondt J, Demuynck K, et al. Ecapa-tdnn: emphasized channel attention, propagation and aggregation in tdnn based speaker verification[C]∥Interspeech, Shanghai, China, 2020: 3830-3834.
16 Shen H, Yang Y, Sun G, et al. Improving fairness in speaker verification via group-adapted fusion network[C]∥ICASSP 2022-2022 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), Singapore, 2022: 7077-7081.
[1] Meng-xue ZHAO,Xiang-jiu CHE,Huan XU,Quan-le LIU. A method for generating proposals of medical image based on prior knowledge optimization [J]. Journal of Jilin University(Engineering and Technology Edition), 2025, 55(2): 722-730.
[2] Yuan-ning LIU,Zi-nan ZANG,Hao ZHANG,Zhen LIU. Deep learning-based method for ribonucleic acid secondary structure prediction [J]. Journal of Jilin University(Engineering and Technology Edition), 2025, 55(1): 297-306.
[3] Xi ZHANG,Shao-ping KU. Facial super-resolution reconstruction method based on generative adversarial networks [J]. Journal of Jilin University(Engineering and Technology Edition), 2025, 55(1): 333-338.
[4] Hui-zhi XU,Shi-sen JIANG,Xiu-qing WANG,Shuang CHEN. Vehicle target detection and ranging in vehicle image based on deep learning [J]. Journal of Jilin University(Engineering and Technology Edition), 2025, 55(1): 185-197.
[5] Lu Li,Jun-qi Song,Ming Zhu,He-qun Tan,Yu-fan Zhou,Chao-qi Sun,Cheng-yu Zhou. Object extraction of yellow catfish based on RGHS image enhancement and improved YOLOv5 network [J]. Journal of Jilin University(Engineering and Technology Edition), 2024, 54(9): 2638-2645.
[6] Lei ZHANG,Jing JIAO,Bo-xin LI,Yan-jie ZHOU. Large capacity semi structured data extraction algorithm combining machine learning and deep learning [J]. Journal of Jilin University(Engineering and Technology Edition), 2024, 54(9): 2631-2637.
[7] Bai-you QIAO,Tong WU,Lu YANG,You-wen JIANG. A text sentiment analysis method based on BiGRU and capsule network [J]. Journal of Jilin University(Engineering and Technology Edition), 2024, 54(7): 2026-2037.
[8] Xin-gang GUO,Ying-chen HE,Chao CHENG. Noise-resistant multistep image super resolution network [J]. Journal of Jilin University(Engineering and Technology Edition), 2024, 54(7): 2063-2071.
[9] Li-ping ZHANG,Bin-yu LIU,Song LI,Zhong-xiao HAO. Trajectory k nearest neighbor query method based on sparse multi-head attention [J]. Journal of Jilin University(Engineering and Technology Edition), 2024, 54(6): 1756-1766.
[10] Ming-hui SUN,Hao XUE,Yu-bo JIN,Wei-dong QU,Gui-he QIN. Video saliency prediction with collective spatio-temporal attention [J]. Journal of Jilin University(Engineering and Technology Edition), 2024, 54(6): 1767-1776.
[11] Yu-kai LU,Shuai-ke YUAN,Shu-sheng XIONG,Shao-peng ZHU,Ning ZHANG. High precision detection system for automotive paint defects [J]. Journal of Jilin University(Engineering and Technology Edition), 2024, 54(5): 1205-1213.
[12] Xiong-fei LI,Zi-xuan SONG,Rui ZHU,Xiao-li ZHANG. Remote sensing change detection model based on multi⁃scale fusion [J]. Journal of Jilin University(Engineering and Technology Edition), 2024, 54(2): 516-523.
[13] Guo-jun YANG,Ya-hui QI,Xiu-ming SHI. Review of bridge crack detection based on digital image technology [J]. Journal of Jilin University(Engineering and Technology Edition), 2024, 54(2): 313-332.
[14] Bin ZHAO,Cheng-dong WU,Xue-jiao ZHANG,Ruo-huai SUN,Yang JIANG. Target grasping network technology of robot manipulator based on attention mechanism [J]. Journal of Jilin University(Engineering and Technology Edition), 2024, 54(12): 3423-3432.
[15] Yong WANG,Yu-xiao BIAN,Xin-chao LI,Chun-ming XU,Gang PENG,Ji-kui WANG. Image dehazing algorithm based on multiscale encoding decoding neural network [J]. Journal of Jilin University(Engineering and Technology Edition), 2024, 54(12): 3626-3636.
Viewed
Full text


Abstract

Cited
  Shared   
  Discussed   
No Suggested Reading articles found!
Copyright © Journal of Jilin University(Engineering and Technology Edition), All Rights Reserved.
Tel: +86-431-85095297　Fax: +86-431-85094074　E-mail: xbgxb@jlu.edu.cn
Powered by Beijing Magtech Co. Ltd