电机是工业生产中常用的动力设备之一，常运行于复杂多变的工况，容易发生故障。转子-轴承系统是电机的关键动力输出部件，一旦发生故障会对设备安全平稳运行造成影响。因此，在变工况下对电机转子-轴承系统开展故障诊断方法研究具有重要意义。振动信号包含了电机丰富的状态信息，变工况下电机的振动信号存在分布差异。信号处理方法需要较强的先验知识且算法开发难度大。经典深度学习方法基于训练和测试数据同分布的假设，难以在变工况下完成准确诊断。本文以电机转子-轴承系统为研究对象，研究其在变工况下的智能故障诊断方法，最终在嵌入式设备上进行部署和验证。

针对数据质量差导致深度学习模型诊断准确率难以保证的问题，在变工况下采用图像对振动信号进行表征，研究了极坐标式和矩阵式振动图像作为样本在智能故障诊断中的表现，包括对称极坐标图像（Symmetrical Polar Coordinates Image，SPCI）、经验模态分解-角度自适应分配的极坐标图像（EMD-Angle Adaptive Distribution of Polar Coordinates Image，EMD-AADPCI）、灰度图像（Gray Image，GI）和灰度纹理图像（Gray Texture Image，GTI）共4种振动图像，通过图像特征分析和诊断准确率对样本生成方法进行衡量，选择适用于变工况电机转子-轴承系统故障诊断的振动图像，为智能故障诊断提供数据基础。

针对变工况数据分布存在差异，经典深度学习网络无法准确诊断且泛化能力差的问题，在变负载工况下进行基于深度迁移学习的智能故障诊断方法研究。将卷积神经网络（Convolutional Neural Networks，CNN）与自适应批归一化（Adaptive Batch Normalization，AdaBN）结合，提出了一种深度迁移学习网络（Deep Transfer Learning Network，DTLN）。DTLN在13种工况的诊断准确率为98.7%，在其中8种未知工况的诊断准确率为99.0%，泛化性能良好，实现了由无载到带载、已知到未知工况的迁移诊断。

针对模型结构庞大、计算复杂度高且训练时间长的问题，在变转速工况下进行了基于轻量化CNN的智能故障诊断方法研究。通过自适应激活函数、分类器结构优化和多尺度特征提取，提出了一种自适应激活的轻量化多尺度卷积神经网络（Adaptively Activated Lightweight Multiscale CNN，Ada-act LMCNN）。Ada-act LMCNN在加速实验中诊断准确率高于95%，加速带载实验中诊断准确率高于98%，混合实验中诊断准确率达到98.5%。其参数量、浮点运算次数（Floating-point Operations Per second，FLOPs）、模型大小和训练时间在对比方法中均为最优，兼具轻量化和诊断准确的优势。

针对智能故障诊断模型应用部署问题，提出了一种轻量化深度迁移学习网络（Lightweight Deep Transfer Learning Network，LDTLN），在包含恒定负载、时变负载、加速和加速带载实验，共20种工况下对模型性能进行了验证。研究了不同批次大小对模型诊断结果的影响，结合诊断准确率和推理时间对超参数进行了设定。基于PyQt框架开发了电机智能故障诊断软件，在Raspberry Pi 4B上进行了软件部署，功能实现正常。

Motor is one of the commonly power equipment in industrial production. It often runs in complex and variable operating conditions, making it prone to failure. Rotor-bearing system is the key power output component of motor. Once a fault occurs, it will affect the safe and stable operation of equipment. Therefore, researching the fault diagnosis method of motor rotor-bearing system under variable conditions is crucial. Vibration signal contains rich state information of motor. The distribution of vibration signal of motor under variable conditions is varies. Signal processing methods require strong prior knowledge and its algorithm development is difficult. Classical deep learning method assume that the training and test data are identically distributed, and it is difficult to complete accurate diagnosis under variable conditions. This thesis focus on the motor rotor-bearing system and researched the intelligent fault diagnosis methods under variable conditions, deployed and verified method on the embedded device.

Aiming at the problem of poor data quality leading to low diagnostic accuracy of deep learning models, images are used to characterize the vibration signal under variable operating conditions, and the performance of polar coordinate and matrix vibration images as samples in intelligent fault diagnosis is researched. Vibration images include four kinds: Symmetric Polar Coordinate Images (SPCI), Empirical Mode Decomposition-Angle Adaptive Distribution of Polar Coordinate Images (EMD-AADPCI), Gray Image (GI) and Gray Texture Image (GTI). The sample generation method is measured by image feature analysis and diagnostic accuracy, and the vibration images which suitable for fault diagnosis of motor rotor-bearing system under variable conditions are selected to provide data basis for intelligent fault diagnosis.

Aiming at the issue that the data distribution of variable operating conditions is different, the classical deep learning network cannot accurately diagnose and the generalization ability is poor. The intelligent fault diagnosis method based on deep transfer learning is researched under variable load conditions. Combining Convolutional Neural Networks (CNN) with Adaptive Batch Normalization (AdaBN), the Deep Transfer Learning Network (DTLN) fault diagnosis method is proposed. The accuracy of DTLN is 98.7% in 13 conditions and 99.0% in 8 unknown conditions. The generalization performance is good, and the transfer diagnosis from no-load to on-load and known to unknown condition is realized.

Aiming at the issue of large model structure, high computational complexity and long training time, the intelligent fault diagnosis method based on lightweight CNN is researched under variable speed conditions. The Adaptively Activated Lightweight Multiscale CNN (Ada-act LMCNN) is developed by adaptive activation function, classifier structure optimization and multi-scale feature extraction. The accuracy of Ada-act LMCNN is higher than 95% in the accelerated test, higher than 98% in the accelerated load test, and 98.5% in the mixed case. The parameters, Floating-point Operations Per second (FLOPs), model size and training time are all the best in the comparison method, which has the advantages of lightweight and accurate diagnosis.

Aiming at the application deployment of intelligent fault diagnosis method, the Lightweight Deep Transfer Learning Network (LDTLN) is proposed. The model performance is verified under 20 conditions including constant load, time-varying load, accelerated and accelerated load test. The influence of different batch sizes on the diagnosis results of the model is researched, and the hyperparameters are set according to the diagnostic accuracy and test time. The motor intelligent fault diagnosis software is developed based on the PyQt framework. The software is deployed on Raspberry Pi 4B, and its function is realized normally.

[1] Liu R N, Yang B Y, Zio E, et al. Artificial intelligence for fault diagnosis of rotating machinery: A review[J]. Mechanical Systems and Signal Processing, 2018, 108: 33-47.

[2] 陈伟. 基于深度迁移学习的异步电机故障诊断方法研究[D]. 徐州: 中国矿业大学, 2022.

[3] Yan R Q, Shang Z G, Xu H, et al. Wavelet transform for rotary machine fault diagnosis:10 years revisited[J]. Mechanical Systems and Signal Processing, 2023, 200: 110545.

[4] Lei Y G, Lin J, He Z J, et al. A review on empirical mode decomposition in fault diagnosis of rotating machinery[J]. Mechanical Systems and Signal Processing, 2013, 35: 108-126.

[5] Amarnath M, Sugumaran V, Kumar H. Exploiting sound signals for fault diagnosis of bearings using decision tree[J]. Measurement, 2013, 46(3): 1250-1256.

[6] Widodo A, Yang B S. Support vector machine in machine condition monitoring and fault diagnosis[J]. Mechanical Systems and Signal Processing, 2007, 21(6): 2560-2574.

[7] Saravanan N, Siddabattuni V N S K, Ramachandran K I. Fault diagnosis of spur bevel gear box using artificial neural network (ANN), and proximal support vector machine (PSVM)[J]. Applied Soft Computing, 2010, 10(1): 344-360.

[8] Wang D C, Guo Q W, Song Y, et al. Application of multiscale learning neural network based on CNN in bearing fault diagnosis[J]. Journal of Signal Processing Systems, 2019, 91(10): 1205-1217.

[9] Zhu J J, Jiang Q S, Shen Y H, et al. Application of recurrent neural network to mechanical fault diagnosis: a review[J]. Journal of Mechanical Science and Technology, 2022, 36(2): 527-542.

[10] Long Z, Zhang X F, Zhang L, et al. Motor fault diagnosis using attention mechanism and improved AdaBoost driven by multi-sensor information[J]. Measurement, 2021, 170: 108718.

[11] 周兴康, 余建波. 基于深度一维残差卷积自编码网络的齿轮箱故障诊断[J]. 机械工程学报, 2020, 56(7): 96-108.

[12] Zou F Q, Zhang H F, Sang S T, et al. An anti-noise one-dimension convolutional neural network learning model applying on bearing fault diagnosis[J]. Measurement, 2021, 186: 110236.

[13] 王崇宇, 郑召利, 刘天源, 等. 基于卷积神经网络的汽轮机转子不平衡与不对中故障检测方法研究[J]. 中国电机工程学报, 2021, 41(7): 2417-2427.

[14] Junior R F R, Santos Areias I A, Campos M M, et al. Fault detection and diagnosis in electric motors using 1d convolutional neural networks with multi-channel vibration signals[J]. Measurement, 2022, 190: 110759.

[15] 杨洁, 万安平, 王景霖, 等. 基于多传感器融合卷积神经网络的航空发动机轴承故障诊断[J]. 中国电机工程学报, 2022, 42(13): 4933-4942.

[16] 胡茑庆, 陈徽鹏, 程哲, 等. 基于经验模态分解和深度卷积神经网络的行星齿轮箱故障诊断方法[J]. 机械工程学报, 2019, 55(7): 9-18.

[17] 张鑫, 郭顺生, 李益兵, 等. 基于拉普拉斯特征映射和深度置信网络的半监督故障识别[J]. 机械工程学报, 2020, 56(1): 69-81.

[18] Xie J S, Li Z Y, Zhou Z T, et al. A novel bearing fault classification method based on XGBoost: The fusion of deep learning-based features and empirical features[J]. IEEE Transactions on Instrumentation and Measurement, 2020, 70: 1-9.

[19] 王骁贤, 陆思良, 何清波, 等. 变转速工况下基于多传感器信号深度特征融合的电机故障诊断研究[J]. 仪器仪表学报, 2022, 43(3): 59-67.

[20] 王振亚, 刘韬, 伍星. 互信息规范的卷积神经网络及其在轴承故障特征提取中的应用[J/OL]. 机械工程学报, 1-10[2024-03-11]. http://kns.cnki.net/kcms/detail/11.2187.TH.20231018.1101.002.html.

[21] Fan H W, Ma J T, Zhang X H, et al. Intelligent data expansion approach of vibration gray texture images of rolling bearing based on improved WGAN-GP[J]. Advances in Mechanical Engineering, 2022, 14(3): 1-11.

[22] Zhao J, Yang S P, Li Q, et al. A new bearing fault diagnosis method based on signal-to-image mapping and convolutional neural network[J]. Measurement, 2021, 176: 109088.

[23] 宫文峰, 陈辉, 张泽辉, 等. 基于改进卷积神经网络的滚动轴承智能故障诊断研究[J]. 振动工程学报, 2020, 33(2): 400-413.

[24] 宋浏阳, 李石, 王芃鑫, 等. 基于动态统计滤波与深度学习的智能故障诊断方法[J]. 仪器仪表学报, 2019, 40(7): 39-46.

[25] Lyu P, Zhang K W, Yu W B, et al. A novel RSG-based intelligent bearing fault diagnosis method for motors in high-noise industrial environment[J]. Advanced Engineering Informatics, 2022, 52: 101564.

[26] 康守强, 刘哲, 王玉静, 等. 基于改进DQN网络的滚动轴承故障诊断方法[J]. 仪器仪表学报, 2021, 42(3): 201-212.

[27] 余晓霞, 汤宝平, 魏静, 等. 强背景噪声条件下自适应图卷积神经网络的航空发动机附件机匣故障诊断方法[J]. 仪器仪表学报, 2021, 41(8): 78-86.

[28] 孙曙光, 张婷婷, 王景芹, 等. 基于连续小波变换和MTL-SEResNet的断路器故障程度评估[J]. 仪器仪表学报, 2022, 43(6): 162-173.

[29] Tran M Q, Liu M K, Tran Q V, et al. Effective fault diagnosis based on wavelet and convolutional attention neural network for induction motors[J]. IEEE Transactions on Instrumentation and Measurement, 2021, 71: 1-13.

[30] 朱霄珣, 罗学智, 叶行飞, 等. 基于深度特征学习的汽轮机转子状态识别方法[J]. 中国电机工程学报, 2021, 41(2): 432-442.

[31] Zhang W W, Chen D J, Kong Y. Self-supervised joint learning fault diagnosis method based on three-channel vibration images[J]. Sensors, 2021, 21(14): 4774.

[32] Cao X G, Xu X, Duan Y, et al. Health status recognition of rotating machinery based on deep residual shrinkage network under time-varying conditions[J]. IEEE Sensors Journal, 2022, 22(19): 18332-18348.

[33] 薛策译. 基于振动图像深度学习的滚动轴承智能故障诊断研究[D]. 西安: 西安科技大学, 2022.

[34] Tang Y, Zhang X F, Huang S, et al. Multi-sensor-driven motor fault diagnosis method based on visual features[J]. IEEE Transactions on Industrial Informatics, 2022, 19(4): 5902-5914.

[35] Abdeljaber O, Sassi S, Avci O, et al. Fault detection and severity identification of ball bearings by online condition monitoring[J]. IEEE Transactions on Industrial Electronics, 2018, 66(10): 8136-8147.

[36] Sun W J, Shao S Y, Zhao R, et al. A sparse auto-encoder-based deep neural network approach for induction motor faults classification[J]. Measurement, 2016, 89: 171-178.

[37] Souza R M, Nascimento E G S, Miranda U A, et al. Deep learning for diagnosis and classification of faults in industrial rotating machinery[J]. Computers & Industrial Engineering, 2021, 153: 107060.

[38] Gao S Z, Xu L T, Zhang Y M, et al. Rolling bearing fault diagnosis based on SSA optimized self-adaptive DBN[J]. ISA Transactions, 2022, 128: 485-502.

[39] Fan H W, Ma N G, Zhang X H, et al. New intelligent fault diagnosis approach of rolling bearing based on improved vibration gray texture image and vision transformer[J]. Proceedings of the Institution of Mechanical Engineers, Part C: Journal of Mechanical Engineering Science, 2022, Early access, DOI:10.1177/09544062221085871.

[40] 李涛, 段礼祥, 张东宁, 等. 自适应卷积神经网络在旋转机械故障诊断中的应用[J]. 振动与冲击, 2020, 39(16): 275-282+288.

[41] Chen X H, Zhang B K, Gao D. Bearing fault diagnosis base on multi-scale CNN and LSTM model[J]. Journal of Intelligent Manufacturing, 2021, 32(4): 971-987.

[42] Xu Y, Li Z X, Wang S Q, et al. A hybrid deep-learning model for fault diagnosis of rolling bearings[J]. Measurement, 2021, 169: 108502.

[43] Kumar P, Hati A S. Deep convolutional neural network based on adaptive gradient optimizer for fault detection in SCIM[J]. ISA Transactions, 2021, 111: 350-359.

[44] 雷春丽, 夏奔锋, 薛林林, 等. 优化深度残差网络及其在强噪声环境下滚动轴承故障诊断中的应用[J]. 振动工程学报, 2023, 36(6): 1754-1763.

[45] 施健聪, 王兴龙, 张俊. 基于双重注意力机制的异步电机故障诊断方法[J]. 振动与冲击, 2023, 42(21): 110-118.

[46] 文斌, 李知聪, 朱晗, 等. MHSACAE-CNN在噪声下的电机轴承故障诊断[J]. 振动工程学报, 2023, 36(4): 1169-1178.

[47] Ye Z, Yu J B. Deep morphological convolutional network for feature learning of vibration signals and its applications to gearbox fault diagnosis[J]. Mechanical Systems and Signal Processing, 2021, 161: 107984.

[48] Wang H, Liu Z L, Peng D D, et al. Attention-guided joint learning CNN with noise robustness for bearing fault diagnosis and vibration signal denoising[J]. ISA Transactions, 2022, 128: 470-484.

[49] Zhao X L, Yao J Y, Deng W X, et al. Intelligent fault diagnosis of gearbox under variable working conditions with adaptive intraclass and interclass convolutional neural network[J]. IEEE Transactions on Neural Networks and Learning Systems, 2023, 34(9): 6339-6353.

[50] Zhang W, Peng G L, Li C H, et al. A new deep learning model for fault diagnosis with good anti-noise and domain adaptation ability on raw vibration signals[J]. Sensors, 2017, 17(2): 425.

[51] 沈长青, 汤盛浩, 江星星, 等. 独立自适应学习率优化深度信念网络在轴承故障诊断中的应用研究[J]. 机械工程学报, 2019, 55(7): 81-88.

[52] 王金瑞, 李舜酩, 钱巍巍, 等. 转速大波动下滚动轴承智能故障诊断方法研究[J]. 振动工程学报, 2020, 33(2): 391-399.

[53] An Z H, Li S M, Wang J R, et al. A novel bearing intelligent fault diagnosis framework under time-varying working conditions using recurrent neural network[J]. ISA Transactions, 2020, 100: 155-170.

[54] Liu R N, Wang F, Yang B Y, et al. Multi-scale kernel based residual convolutional neural network for motor fault diagnosis under nonstationary conditions[J]. IEEE Transactions on Industrial Informatics, 2019, 16(6): 3797-3806.

[55] 叶楠, 常佩泽, 张露予, 等. 基于改进后半监督深度信念网络的多工况轴承故障诊断研究[J]. 机械工程学报, 2021, 57(15): 80-90.

[56] Zhao X L, Yao J Y, Deng W X, et al. Multiscale deep graph convolutional networks for intelligent fault diagnosis of rotor-bearing system under fluctuating working conditions[J]. IEEE Transactions on Industrial Informatics, 2022, 19(1): 166-176.

[57] 李巍华, 何琛, 陈祝云, 等. 基于对称式对比学习的齿轮箱无监督故障诊断方法[J]. 仪器仪表学报, 2022, 43(3): 121-131.

[58] 赵小强, 柴靖轩. 改进卷积胶囊网络的滚动轴承故障诊断方法[J]. 振动工程学报, 2024, 37(5):885-895.

[59] Lu W N, Liang B, Cheng Y, et al. Deep model based domain adaptation for fault diagnosis[J]. IEEE Transactions on Industrial Electronics, 2016, 64(3): 2296-2305.

[60] 雷亚国, 杨彬, 杜兆钧, 等. 大数据下机械装备故障的深度迁移诊断方法[J]. 机械工程学报, 2019, 55(7): 1-8.

[61] Ragab M, Chen Z H, Wu M, et al. Adversarial multiple-target domain adaptation for fault classification[J]. IEEE Transactions on Instrumentation and Measurement, 2020, 70: 1-11.

[62] Shen C Q, Wang X, Wang D, et al. Dynamic joint distribution alignment network for bearing fault diagnosis under variable working conditions[J]. IEEE Transactions on Instrumentation and Measurement, 2021, 70: 1-13.

[63] Han T, Li Y F, Qian M. A hybrid generalization network for intelligent fault diagnosis of rotating machinery under unseen working conditions[J]. IEEE Transactions on Instrumenta-tion and Measurement, 2021, 70: 1-11.

[64] Wang R, Huang W G, Shi J J, et al. A higher-order moment matching based fine-grained adversarial domain adaptation method for intelligent bearing fault diagnosis[J]. Measurement Science and Technology, 2022, 33(5): 055022.

[65] Wu Y C, Zhao R Z, Ma H R, et al. Adversarial domain adaptation convolutional neural network for intelligent recognition of bearing faults[J]. Measurement, 2022, 195: 111150.

[66] 董绍江, 朱朋, 裴雪武, 等. 基于子领域自适应的变工况下滚动轴承故障诊断[J]. 吉林大学学报(工学版), 2022, 52(2): 288-295.

[67] Xiao Y M, Shao H D, Min Z S, et al. Multiscale dilated convolutional subdomain adaptation network with attention for unsupervised fault diagnosis of rotating machinery cross operating conditions[J]. Measurement, 2022, 204: 112146.

[68] 高晗, 田育龙, 许封元, 等. 深度学习模型压缩与加速综述[J]. 软件学报, 2021, 32(1): 68-92.

[69] Iandola F N, Han S, Moskewicz M W, et al. SqueezeNet: AlexNet-level accuracy with 50x fewer parameters and <0.5 MB model size[EB/OL]. (2016–11–04)[2024–03–11]. http://arxiv.org/abs/1602.07360.

[70] Zhang X Y, Zhou X Y, Lin M X, et al. Shufflenet: An extremely efficient convolutional neural network for mobile devices[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. Salt Lake City, UT, USA: IEEE, 2018: 6848-6856.

[71] Howard A G, Zhu M L, Chen B, et al. MobileNets: Efficient convolutional neural networks for mobile vision applications[EB/OL]. (2017-04-16)[2024-03-11]. http://arxiv.org/abs/17

04.04861.

[72] Han K, Wang Y H, Tian Q, et al. GhostNet: More features from cheap opera-tions[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. Seattle, WA, USA: IEEE/CVF, 2020: 1580-1589.

[73] 邓飞跃, 吕浩洋, 顾晓辉, 等. 基于轻量化神经网络Shuffle-SENet的高速动车组轴箱轴承故障诊断方法[J]. 吉林大学学报(工学版), 2022, 52(2): 474-482.

[74] 余志锋, 熊邦书, 李新民, 等. 基于改进的SqueezeNet直升机滚动轴承故障诊断[J]. 航空动力学报, 2022, 37(6): 1162-1170.

[75] Jiménez-Guarneros M, Morales-Perez C, Rangel-Magdaleno J. Diagnostic of combined mechanical and electrical faults in ASD-powered induction motor using MODWT and a lightweight 1-D CNN[J]. IEEE Transactions on Industrial Informatics, 2021, 18(7): 4688-4697.

[76] Xu Y D, Yan X A, Feng K, et al. Attention-based multiscale denoising residual convolu-tional neural networks for fault diagnosis of rotating machinery[J]. Reliability Engineering & System Safety, 2022, 226: 108714.

[77] Zhong H Y, Lv Y, Yuan R, et al. Bearing fault diagnosis using transfer learning and self-attention ensemble lightweight convolutional neural network[J]. Neurocomputing, 2022, 501: 765-777.

[78] 段洁利, 于世伟, 解明坤, 等. 基于一维轻量化CNN的山地索道轴承故障诊断[J]. 农业工程学报, 2023, 39(14): 70-79.

[79] Ma S H, Sun H C, Gao S, et al. A real-time mechanical fault diagnosis approach based on lightweight architecture search considering industrial edge deployments[J]. Engineering Applications of Artificial Intelligence, 2023, 123: 106433.

[80] Lu S L, Qian G, He Q B, et al. In situ motor fault diagnosis using enhanced convolutional neural network in an embedded system[J]. IEEE Sensors Journal, 2019, 20(15): 8287-8296.

[81] 吴启航, 丁晓喜, 何清波, 等. 齿轮箱故障边缘智能诊断方法及应用研究[J]. 仪器仪表学报, 2024, 45(1): 70-80.

[82] Chen L, An K, Huang D L, et al. Noise-boosted convolutional neural network for edge-based motor fault diagnosis with limited samples[J]. IEEE Transactions on Industrial Informatics, 2023, 19(9): 9491-9502.

[83] An K, Lu J F, Zhu Q J, et al. Edge solution for real-time motor fault diagnosis based on efficient convolutional neural network[J]. IEEE Transactions on Instrumentation and Measurement, 2023, 72: 1-12.

[84] Lu S L, Lu J F, An K, et al. Edge computing on IoT for machine signal processing and fault diagnosis: A review[J]. IEEE Internet of Things Journal, 2023, 10(13): 11093-11116.

[85] Fan H W, Xue C Y, Ma J T, et al. A novel intelligent diagnosis method of rolling bearing and rotor composite faults based on vibration signal-to-image mapping and CNN-SVM[J]. Measurement Science and Technology, 2023, 34(4): 044008.

[86] 樊红卫, 邵偲洁. 基于灰度图像纹理分析的电主轴不平衡故障诊断方法[J]. 制造技术与机床, 2019(11): 130-134.

[87] 邱锡鹏. 神经网络与深度学习[M]. 北京: 机械工业出版社, 2020: 119.

[88] 马嘉腾. 不平衡小样本下基于深度学习的滚动轴承故障诊断方法研究[D]. 西安: 西安科技大学, 2023.

[89] Hendrycks D, Gimpel K. Gaussian error linear units (GELUs)[EB/OL]. (2023-06-05)[2024-03-11]. http://arxiv.org/abs/1606.08415.

[90] Li Y H, Wang N Y, Shi J P, et al. Adaptive batch normalization for practical domain adaptation[J]. Pattern Recognition, 2018, 80: 109-117.

[91] LeCun Y, Bottou L, Bengio Y, et al. Gradient-based learning applied to document recognition[J]. Proceedings of the IEEE, 1998, 86(11): 2278-2324.

[92] Krizhevsky A, Sutskever I, Hinton G E. ImageNet classification with deep convolutional neural networks[J]. Communications of the ACM, 2017, 60(6): 84-90.

[93] Ma N N, Zhang X Y, Liu M, et al. Activate or not: Learning customized active-tion[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. Nashville, TN, USA: IEEE/CVF, 2021: 8032-8042.

[94] Chen J R, Kao S, He H, et al. Run, Don't walk: Chasing higher FLOPS for faster neural networks[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. Vancouver, BC, CAN: IEEE/CVF, 2023: 12021-12031.

[95] Yu W H, Zhou P, Yan S C, et al. Inceptionnext: When inception meets convnext[C]//Pro-ceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. Seattle, WA, USA: IEEE/CVF, 2024: 5672-5683.

[96] Szegedy C, Liu W, Jia Y Q, et al. Going deeper with convolutions[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. Boston, MA, USA: IEEE, 2015: 1-9.