Research on Voiceprint Recognition based on Densenet
DOI:
https://doi.org/10.54097/71cfa872Keywords:
Voiceprint Recognition, Depth Separable Convolutional, FBank, DenseNet, BiGRUAbstract
To date, most methods for voiceprint recognition are implemented based on convolutional neural networks. In order to combine the spatial features and temporal features of sound and improve computational efficiency, a combined model of DenseNet169 and Bidirectional Gated Recurrent Unit (BiGRU), named DenseNet169-BiGRU, is proposed. Firstly, deep separable convolution is employed to reduce the number of network parameters and enhance computational efficiency. The speech FBank feature map is extracted after preprocessing, and experimental results show that the DenseNet169-BiGRU model outperforms both DenseNet169 and DenseNet169-GRU models, with an improvement in accuracy rate of 2.7% compared to DenseNet169, and an improvement of 1.1% compared to Densenet-GRU. This validates the effectiveness of the proposed method in both improving computational efficiency and achieving good performance.
Downloads
References
Chen Guo-guo,Chen Jia-yu,Na Xing-yu’et al. Kaldi speech recognition combat[M]. 1st ed. Beijing:Publishing House of Electronics Industry 2020. (in Chinese)
HANSEN J H L, HASAN T. Speaker Recognition by Machines and Humans: A tutorial review[J/OL]. IEEE Signal Processing Magazine, 2015, 32(6): 74-99. DOI:10.1109/ MSP. 2015. 246 2851.
LIU D, XU J, ZHANG P, et al. A unified system for multilingual speech recognition and language identification [J/OL]. Speech Communication, 2021, 127: 17-28. DOI: 10. 1016/ j.specom.2020.12.008.
ABLIMIT M, XUELI M, HAMDULLA A. Language Identification Research Based on Dual Attention Mechanism[C/OL]//2021 IEEE 2nd International Conference on Pattern Recognition and Machine Learning (PRML). 2021: 241-246[2023-12-10]. https:// ieeexplore. ieee.org/ document/ 9520699. DOI:10.1109/PRML52754.2021.9520699.
ZHANG C, KOISHIDA K, HANSEN J. Text-Independent Speaker Verification Based on Triplet Convolutional Neural Network Embeddings[J/OL]. IEEE/ACM Transactions on Audio, Speech, and Language Processing, 2018, 26: 1-1. DOI:10.1109/TASLP.2018.2831456.
SARKAR A, TAN Z H. Text Dependent Speaker Verification Using un-supervised HMM-UBM and Temporal GMM-UBM[C/OL]. 2016. DOI:10.21437/Interspeech.2016-362.
HAO M, LIU H, LI Y, et al. Speech Tracking Based on Cluster Analysis and Speaker Recognition[J/OL]. Computer and Modernization, 2020, 4: 11-17. DOI:10.3969/j.issn.1006-2475. 2020.04.002.
LIU Z, WU Z, LI T, et al. GMM and CNN Hybrid Method for Short Utterance Speaker Recognition[J/OL]. IEEE Transactions on Industrial Informatics, 2018, PP: 1-1. DOI: 10. 1109/ TII.2018.2799928.
LUO K, FU L. Research and application of voiceprint recognition based on a deep recurrent neural network[M/OL]. 2019: 309-316. DOI:10.1201/9780429468605-43.
HE Y, YINGYAN D, PENG W, et al. VOICEPRINT RECOGNITION BASED ON CNN-LSTM NETWORK [J/OL]. Computer Applications and Software, 2019 [2023-12-15]. http://en.cnki. com.cn/Article_ en/CJFD Total-JYRJ 2019 04027.htm.
MA Y, YUAN M, QI C, et al. Reacher of Feature Extratcion from spectrogram Based on Pulse Coupled Neural Network in Speaker Recognition[J/OL]. 2005(20)[2023-12-11]. https:// ir. lzu. edu.cn//handle/262010/127329. DOI:10/127329.
CUI L, WANG Z. Study on Voiceprint Recognition Based on Mixed Features of LFBank and FBank[J]. Computer Science, 2022, 49(S2): 621-625.
CHU X, YANG H, YAN D, et al. Research on CNN-LSTM Speaker Recognition Based on Global Attention Mechanism [J/OL]. Audio Engineering, 2022, 46(2): 38-45. DOI:10. 16311/ j.audioe.2022.02.009.
RONG Y, FANG Y, TIAN P, et al. Voiceprint recognition based on knowledge distillation and ResNet[J]. Journal of Chongqing University, 2023, 46(1): 113-124.
WANG F, XU Y. LSTM-Based End-to-End Voiceprint Recognition Algorithm Implementation[J/OL]. Software Engineering and Applications, 2021, 10: 467. DOI:10. 12677/ SEA. 2021.104052.
WANG ZHI-YUE C L. End to End Voiceprint Recgonition Based on Nonlinear Stacked Bidirectional Network[J]. Computer and Modernization, 2022, 0(03): 13.
ZHANG Y, ZHANG Z. Application of DenseNet in voiceprint recognition[J]. Computer Engineering & Science, 2022, 44(1): 132-137.
JIANG G, FU Y. Simulation of User Login Speech Recognition Model Based on Multi-Task Training[J]. Computer Simulation, 2022, 39(9): 190-194.
LIU X, SONG W, CHEN X, et al. BLSTM-CTC SPEECH RECOGNITION BASED ON MULTI-CORE CONVOLUTIONAL FUSION NETWORK[J]. Computer Applications and Software, 2021, 38(11): 167-173.
Downloads
Published
Issue
Section
License
Copyright (c) 2023 Frontiers in Computing and Intelligent Systems
This work is licensed under a Creative Commons Attribution 4.0 International License.
