To accurately identify the abnormal door-closing noise, we propose a method to recognize the time-frequency image of door closing sound based on a multi-scale feature fusion network model. The door-closing sound signal is transformed into a time-frequency image through wavelet analysis, and a classification model based on multi-scale feature fusion is designed. The model introduces multi-scale filters and dilated convolution and adds two improved inception modules to keep the model lightweight. At the same time, richer spatial features can be obtained. The features of different scales are spliced and input to the fully connected layer, and a dropout layer is added to the fully connected layer to suppress overfitting. By comparing the loss and accuracy rate, adjusting different hyperparameters, the optimal model is obtained. The experimental results show that the multi-scale feature fusion network model has a higher accuracy rate than the transfer learning model. Test accuracy rate is 86% and can effectively recognize abnormal door-closing noise. It provides a feasible theoretical basis for the direction of abnormal door noise recognition.
MalenDEScottRA. Improving automobile door-closing sound for customer preference. Noise Control Eng J1993; 41(1): 261–271.
2.
TianYHongliangLSuL, et al. Application of acoustic positioning technology to improve the design of car door closing sound quality. Sci Technol Eng2017; 17(13): 286–290.
3.
GuCFangYDongW. Research on the automatic recognition method of abnormal noise in the car. Automobile Digest2019; 10: 21–25.
4.
KrizhevskyASutskeverIHintonGE.ImageNet classification with deep convolutional neural networks. Adv Neural Inf Process Syst2012; 25(2): 1097–1105.
5.
SimonyanKZissermanA.Very deep convolutional networks for large-scale image recognition. arXiv preprint arXiv: 1409.1556, 2014.
6.
SzegedyCLiuWJiaYet al. Going deeper with convolutions. In: 2015 IEEE Conference on Computer Vision and Pattern Recognition, Boston, MA, 7–12 June 2015, pp. 1–9. New York, IEEE.
7.
HeKZhangXRenSet al. Deep Residual Learning for Image Recognition. In: 2016 IEEE Conference on Computer Vision and Pattern Recognition, Las Vegas, USA, 27–30 June 2016, pp. 770–778. New York, IEEE.
8.
SongWLiSFangL, et al. Hyperspectral image classification with deep feature fusion network. IEEE Trans Geosci Remote Sens2018; 56(6): 3173–3184.
9.
WangWDouSJiangZ, et al. A fast dense spectral–spatial convolution network framework for hyperspectral images classification. Remote Sens2018; 10(7): 1068.
10.
SzegedyCLiuWJiaY Q, et al. Going deeper with convolutions. In: 2015 IEEE conference on computer vision and pattern recognition. Piscataway: IEEE Press, 2015, pp.1–9.
11.
LeeHKwonH. Going deeper with contextual CNN for hyperspectral image classification. IEEE Trans Image Process2017; 26(10): 4843–4855.
12.
ZhangCLiGDuS. Multi-scale dense networks for hyperspectral remote sensing Image Classification. IEEE Trans Geosci Remote Sens2019; 57(11): 9201–9222.
13.
LeiLYanGYuhuiL, et al. Research on the quality evaluation method of vehicle door close sound. J Automot Eng2012; 2(02): 129–133.
14.
SalamonJBelloJP. Deep convolutional neural networks and data augmentation for Environmental Sound Classification. IEEE Signal Process Lett2017; 24(3): 279–283.
15.
DanLiCDWangSWangH, et al. Multi-subject data augmentation for target subject semantic decoding with deep multi-view adversarial learning. Inform Sci2021; 547: 1025–1044.
16.
LalithaSGuptaDZakariahM, et al. Investigation of multilingual and mixed-lingual emotion recognition using enhanced cues with data augmentation. Appl Acoust2020; 170:170.
17.
WanXLYangDPWangYSet al. Time-domain signal reconstruction of vehicle interior noise based on deep learning and compressed sensing techniques. Mech Syst Signal Process2020; 139.
18.
WenningWAifengLIPengLAN. A small sample hyperspectral classification method based on data enhancement. Shandong Province: CN112784774A, 2021.
19.
WangDZhangTZhuR, et al. A low-quality image classification and enhancement method based on model fusion and data enhancement. Shanghai: CN111815529A, 2020. pp.10.
20.
HuangZQiuSLiJ, et al. Road traffic sign identification in weak illumination for intelligent vehicle based on machine vision. Recent Pat Mech Eng2018; 11(2): 127–134.
21.
HongSRongzeXXiaopengG, et al. Off-line Chinese signature verification using convolutional neural network. Recent Pat Eng2019; 13(4): 348–355.
22.
WeiYXiaoHShiH, et al. Revisiting dilated convolution: A simple approach for weakly- and semi-supervised semantic segmentation. In: 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Salt Lake, Utah, 18–23 June 2018, pp. 7268–7277. New York, IEEE.
23.
SercuTGoelV.Dense prediction on sequences with time-dilated convolutions for speech recognition. arXiv preprint arXiv: 1611.09288, 2016.
24.
ZhangQCuiZNiuXet al. Image segmentation with pyramid dilated convolution based on ResNet and U-Net. Lect Notes Comput Sci2017; 10635: 364–372.
ZhangPMengJLuanY, et al. Plant miRNA–lncRNA interaction prediction with the ensemble of CNN and IndRNN. Interdiscip Sci Comput Life Sci2020; 12(1): 82–89.
27.
HuangLJiangHLiS, et al. Two stage residual CNN for texture denoising and structure enhancement on low dose CT image. Comp Methods Programs Biomed2020; 184: 105115.1–105115.11.
28.
ChenWJiangMZhangWGet al. A novel graph convolutional feature based convolutional neural network for stock trend prediction. Inf Sci2021; 556(3): 67–94.
29.
LongRZhangJSunZ, et al. A combined load forecasting of electric vehicles based on freshness and cross-entropy algorithm. Tianjin City: CN111709568A, 2020.
30.
BaoyuFWangLI. A picture recognition method, device, equipment and readable storage medium. Jiangsu Province: CN113191461A, 2021.
31.
XuWJiangLLiC.Improving data and model quality in crowdsourcing using cross-entropy-based noise correction. Inf Sci2021; 546(1): 803–814.
32.
HegdeSMundadaMR. Optimized deep neural network based predictive model for customer attrition analysis in the banking sector. Recent Pat Eng2021; 14(3): 412–421.
33.
ChunqiaoMQingyouDXiaoningP, et al. Predicting student study failure risk using artificial neural network method. Recent Pat Eng2019; 13(4): 382–386.
34.
KingmaDBaJ.Adam: A method for stochastic optimization. arXiv preprint arXiv: 1412.6980, 2014.
35.
WangXWangSCaoJ, et al. Data-driven based tiny-YOLOv3 method for front vehicle detection inducing SPP-net. IEEE Access2020; 8:110227–110236.
36.
KrizhevskyASutskeverIHintonGE.Image net classification with deep convolutional neural networks. Commun ACM2017; 60(6): 84–90.
37.
HanDLiuQFanW. A new image classification method using CNN transfer learning and web data augmentation. Expert Syst Appl2018; 95(2): 43–56.
38.
PanSJYangQ. A survey on transfer learning. IEEE Trans Knowl Data Eng2010; 22(10): 1345–1359.
