Please wait a minute...
浙江大学学报(工学版)
浙江大学学报(工学版)  2024, Vol. 58 Issue (5): 879-890    DOI: 10.3785/j.issn.1008-973X.2024.05.001
计算机技术、通信技术     
基于深度学习的EEG数据分析技术综述
钟博(),王鹏飞,王乙乔,王晓玲*()
华东师范大学 计算机科学与技术学院，上海 200062
Survey of deep learning based EEG data analysis technology
Bo ZHONG(),Pengfei WANG,Yiqiao WANG,Xiaoling WANG*()
School of Computer Science and Technology, East China Normal University, Shanghai 200062, China
 全文: PDF(690 KB)   HTML
摘要：

对近年来的相关工作进行全面分析、横向比较，梳理出基于深度学习的EEG数据分析闭环流程. 对EEG数据进行介绍，从深度学习在EEG数据预处理、特征提取以及模型泛化3个关键阶段的应用进行展开，梳理深度学习算法在相应阶段提供的研究思路和解决方案，包括各阶段所存在的难点与问题. 全方位总结出不同算法的主要贡献和局限性，讨论深度学习技术在各个阶段处理EEG数据时所面临的挑战及未来的发展方向.
关键词： 头皮脑电(EEG);  闭环流程;  深度学习;  预处理;  特征提取;  模型泛化    
Abstract:

A thorough analysis and cross-comparison of recent relevant works was provided, outlining a closed-loop process for EEG data analysis based on deep learning. EEG data were introduced, and the application of deep learning in three key stages: preprocessing, feature extraction, and model generalization was unfolded. The research ideas and solutions provided by deep learning algorithms in the respective stages were delineated, including the challenges and issues encountered at each stage. The main contributions and limitations of different algorithms were comprehensively summarized. The challenges faced and future directions of deep learning technology in handling EEG data at each stage were discussed.
Key words: electroencephalography (EEG)    closed-loop process    deep learning    preprocessing    feature extraction    model generalization
收稿日期: 2023-10-23 出版日期: 2024-04-26
CLC:  TP 392  
基金资助: 国家自然科学基金资助项目（61972155）.
通讯作者: 王晓玲     E-mail: bzhong@stu.ecnu.edu.cn;xlwang@cs.ecnu.edu.cn
作者简介: 钟博（1999—），男，硕士生，从事时间序列、医疗信号建模的研究. orcid.org/0009-0006-2742-1683.E-mail：bzhong@stu.ecnu.edu.cn
服务　　
把本文推荐给朋友
加入引用管理器
E-mail Alert
作者相关文章　　
钟博
王鹏飞
王乙乔
王晓玲

引用本文:

钟博,王鹏飞,王乙乔,王晓玲. 基于深度学习的EEG数据分析技术综述[J]. 浙江大学学报(工学版), 2024, 58(5): 879-890.

Bo ZHONG,Pengfei WANG,Yiqiao WANG,Xiaoling WANG. Survey of deep learning based EEG data analysis technology. Journal of ZheJiang University (Engineering Science), 2024, 58(5): 879-890.

链接本文:

https://www.zjujournals.com/eng/CN/10.3785/j.issn.1008-973X.2024.05.001    或    https://www.zjujournals.com/eng/CN/Y2024/V58/I5/879

图 1  EEG闭环分析流程
频段频率/Hz常见场景
δ<4成人慢波睡眠
θ4~7青少年或意志受挫的成人
α8~15放松、闭眼静息
β16~31警惕或激动
γ>32感知专注
μ8~12睡眠
表 1  EEG频段特征
名称文献主要贡献网络架构
CNNLSTMGAN
EEGdenoiseNet文献[12]通过构造干净EEG和带有EOG和EMG伪影的数据集，以端到端的方式进行训练，
对EOG和EMG伪影进行去噪.		√√—
DenoisingEEG文献[13]
基于GAN，生成器从有噪声的EEG训练数据中进行采样去噪，并将其与
相应的干净EEG信号输入鉴别器中进行比较.		√—√
Auto-Denoising文献[14]
引入样本熵和基于能量阈值的数据归一化方法，将图像恢复的思想应用于脑电信号去噪，
使GAN模型中的生成器能够生成平稳的EEG信号.		—√√
U-NET文献[15]
将EEG信号转换为图像，使用计算机视觉领域分割经典模型U-NET.√——
EEGANet文献[16]
针对EOG伪影，不需要专家进行视觉检查或额外的EOG通道，可用于各种
条件下的EEG信号（无眼动、水平眼动、垂直眼动和眨眼）去伪迹.		√—√
Novel CNN文献[17]
引入多种网络模块，下采样EEG数据减少参数量、增加特征图数量、
提高网络深度，进而提高特征维度.		√——
1D-ResCNN文献[18]
将使用不同尺度卷积核的残差块组合起来，让模型捕捉更丰富的特征，
使得降噪后 EEG 的非线性特征得到显著保持，提高了在未知噪声下的 EEG 降噪性能.		√——
表 2  基于深度学习的EEG信号去噪方法
类型文献

主要贡献
数据增强文献[19]
利用额外的可变伸缩系数，动态调节WGAN-GP中的梯度惩罚项.
文献[20]
使用WGAN从低分辨率的记录中生成高空间分辨率EEG数据.
文献[21]
在WGAN的基础上，提出时空频率-均方误差损失，有助于重构出更具判别性的信号.
数据转换文献[22]
提出基于自编码器改进的深度学习模型，实现由EEG到同步SEEG信号的生成.
文献[23]
提出两阶段匹配策略，确定一对一电极匹配关系，基于CGAN实现EEG到SEEG的生成.
表 3  基于深度学习的EEG信号生成方法
文献主要贡献网络架构
LSTMCNNAEGAN
文献[24]
提取出脑电图信号分析中最常用的各种统计特征作为深度模型LSTM的输入，证明了LSTM是预测癫痫发作的理想工具.√———
文献[25]
在AE模型的基础上引入门控机制对输入向量进行部分遮蔽，促使编码层学习变量之间的关系，挖掘EEG序列内部不同变量间的内在关联，在重建恢复损坏的EEG数据方面具有优越的性能.——√—
文献[26]
通过CNN和AE的组合，捕获维度更低、信息量比例更大的EEG信号表示，适用于带宽有限、延时容忍度低的实时数据传输场景.—√√—
文献[27]
通过CNN提取EEG中蕴含的认知领域知识，对视觉刺激隐向量进行约束，在提高图像分类性能的同时，实现了由EEG图像认知特征生成对应的视觉图像的功能.—√—√
文献[28]
介绍并比较2种专门设计的自编码器学习EEG特征的策略，通道级自编码器专注于每个通道中的特征；图像级自编码器从整个片段中学习特征.—√√—
文献[29]
将EEG数据转化为一连串连续的保留拓扑结构的多光谱图像，保留EEG的空间、光谱和时间结构，从图像序列中学习信号的鲁棒表示，在心理负荷分类任务中表现
出优势.		√√——
表 4  基于深度学习的EEG时域特征提取方法
文献频域分解方式
特征提取/融合方式
文献[30]
提取$ \delta \mathrm{、}\theta \mathrm{、}\alpha \mathrm{、}\beta \mathrm{、}\gamma $波段将每个波段对应的特征向量直接连接
文献[31]
提取$ \delta \mathrm{、}\theta \mathrm{、}\alpha \mathrm{、}\beta $波段将多个子波段堆叠输入网络
文献[32]
提取$ \delta \mathrm{、}\theta \mathrm{、}\alpha \mathrm{、}\beta \mathrm{、}\gamma $波段，将相邻的波段结合形成不同尺度采用自适应的权重学习方法，对不同尺度的分类结果进行融合
文献[33]
提取$ \delta \mathrm{、}\theta \mathrm{、}\alpha \mathrm{、}\beta \mathrm{、}\gamma $波段，为每个受试者自适应地挑选最优波段将多个波段的输出直接连接
文献[34]
提取$ \theta \mathrm{、}\alpha \mathrm{、}\beta $波段将每个波段视为RGB中的一个通道，使用基于CNN的自编码器进行特征提取
文献[35]
将8~30 Hz分解至10个子波段，每个子波段宽为4 Hz，且相邻子波段之间的重叠频带为2 Hz将输入视为图像格式，沿着频率轴滑动卷积核
文献[36]
捕获频率≥2 Hz的信息为每个特定频率学习多个空间滤波器
文献[37]
—使用由小波核启发的时频滤波器
文献[38]
—将复杂网络表示扩展至频域，使用序列卷积从连续频率中学
习特征
表 5  基于深度学习的EEG频域特征提取方法
模型文献

主要贡献使用技术
注意力
机制		信道
建模		空间
信息		频域
信息		时域
信息
SPDNet RNN文献[39]
提出EEG协方差矩阵的时空融合方法，利用整个EEG信号的协方差矩阵提取时空特征，通过EEG通道之间的时空变化来研究驾驶员疲劳检测.——√—√
RACNN文献[40]
分别对不同脑功能区的频谱特征进行卷积计算学习EEG频谱-时空特征，利用注意力机制加权聚合各脑功能区特征，鼓励对最重要的区域赋予更大的注意力权重，揭示不同脑区EEG数据与任务存在的动态相关性.√——√√
RGNN文献[41]
考虑不同脑区之间的不同生物拓扑结构，捕捉EEG信道之间的局部和全局关系，采用符合生物原理的稀疏邻接矩阵捕捉局部和全局EEG通道间的关系. 实验结果表明，对于情感识别，前额叶、枕叶和顶叶可能是信息量最大的区域.—√———
IAG文献[44]
采用额外分支融合空间信息和频率信息，自适应表征不同EEG通道间的内在动态关系，为EEG情绪识别加入更多的特征.—√√√—
MD-AGCN文献[45]
纳入来自时间域、频域和脑功能连接的脑电图信息，将EEG的时域和频域与通道的拓扑结构相结合，以自适应的方式学习与情绪相关的大脑功能连接—√—√√
表 6  基于深度学习的EEG空间特征提取方法
训练目标文献
主要贡献使用技术
自监督
训练		域对抗
训练		数据
增强		多任务
学习		多视图
融合		知识
蒸馏
跨受试者文献[48]
构建片段级别的EEG特征序列，引入自监督任务，整个体系结构
都可以被微调到各种下游脑机接口和EEG分类任务.		√—————
跨受试者文献[50]
引入对抗性训练的方式，减少个体差异的影响，提取具有域
不变性的EEG信号特征.		—√————
跨任务文献[51]
通过聚合空间拼接、频率拼接多任务的学习，避免模型在单个
下游任务上的过拟合.		√——√——
跨视图文献[52]
提出EEG的多视图增强策略，在训练过程中利用多视图的互补信息.√—√—√—
跨模态文献[54]
采用跨模态的知识蒸馏策略，减小头皮EEG和基于耳朵EEG的
睡眠分级之间的性能差距.		—————√
跨数据集文献[55]
提出时频一致性策略，促使基于时间的表示和基于频率的
表示彼此接近.		√—√—√—
表 7  基于深度学习的EEG模型泛化方法
1 HOSSEINU M P, HOSSEINI A, AHI K A review on machine learning for EEG signal processing in bioengineering[J]. IEEE Reviews in Biomedical Engineering, 2020, 14: 204- 218
2 JIANG X, BIAN G B, TIAN Z Removal of artifacts from EEG signals: a review[J]. Sensors, 2019, 19 (5): 987
doi: 10.3390/s19050987
3 ZHANG X, YAO L, WANG X, et al. A survey on deep learning based brain computer interface: recent advances and new frontiers [EB/OL]. [2019-05-10]. https://arxiv.org/pdf/1905.04149.
4 OKAMOTO M, DAN H, SAKAMOTO K, et al Three-dimensional probabilistic anatomical cranio-cerebral correlation via the international 10–20 system oriented for transcranial functional brain mapping[J]. Neuroimage, 2004, 21 (1): 99- 111
doi: 10.1016/j.neuroimage.2003.08.026
5 O'REILLY C, GOSSELIN N, CARRIER J, et al Montreal archive of sleep studies: an open-access resource for instrument benchmarking and exploratory research[J]. Journal of Sleep Research, 2014, 23 (6): 628- 635
doi: 10.1111/jsr.12169
6 GOLDBERGER A L, AMARAL L A N, GLASS L, et al PhysioBank, PhysioToolkit, and PhysioNet: components of a new research resource for complex physiologic signals[J]. Circulation, 2000, 101 (23): e215- e220
7 ZHENG W L, LIU W, LU Y, et al Emotionmeter: a multimodal framework for recognizing human emotions[J]. IEEE Transactions on Cybernetics, 2019, 49 (3): 1110- 1122
doi: 10.1109/TCYB.2018.2797176
8 KOELSTRA S, MUHL C, SOLEYMANI M, et al Deap: a database for emotion analysis; using physiological signals[J]. IEEE Transactions on Affective Computing, 2012, 3 (1): 18- 31
doi: 10.1109/T-AFFC.2011.15
9 ZHAO Q, ZHANG L. Temporal and spatial features of single-trial EEG for brain-computer interface [EB/OL]. [2023-10-01]. https://doi.org/10.1155/2007/37695.
10 MOHANTY R, SETHARES W A, NAIR V A, et al Rethinking measures of functional connectivity via feature extraction[J]. Scientific Reports, 2020, 10 (1): 1298
doi: 10.1038/s41598-020-57915-w
11 SALIS C I, MALISSOYAS A E, BIZOPOULOS P A, et al. Denoising simulated EEG signals: a comparative study of EMD, wavelet transform and Kalman filter [C]// 13th IEEE International Conference on BioInformatics and BioEngineering . Los Alamitos: IEEE, 2013: 1-4.
12 ZHANG H, ZHAO M, WEI C, et al EEGdenoisenet: a benchmark dataset for deep learning solutions of EEG denoising[J]. Journal of Neural Engineering, 2021, 18 (5): 056057
doi: 10.1088/1741-2552/ac2bf8
13 BROPHY E, REDMOND P, FLEURY A, et al Denoising EEG signals for real-world BCI applications using GANs[J]. Frontiers in Neuroergonomics, 2022, 2: 44
14 AN Y, LAM H K, LING S H Auto-denoising for EEG signals using generative adversarial network[J]. Sensors, 2022, 22 (5): 1750
doi: 10.3390/s22051750
15 MASHHADI N, KHUZANI A Z, HEIDARI M, et al. Deep learning denoising for EOG artifacts removal from EEG signals [C]// IEEE Global Humanitarian Technology Conference . Seattle: IEEE, 2020: 1-6.
16 SAWANGJAI P, TRAKULRUANGROJ M, BOONNAG C, et al EEGANet: removal of ocular artifacts from the EEG signal using generative adversarial networks[J]. IEEE Journal of Biomedical and Health Informatics, 2021, 26 (10): 4913- 4924
17 ZHANG H, WEI C, ZHAO M, et al. A novel convolutional neural network model to remove muscle artifacts from EEG [C]// IEEE International Conference on Acoustics, Speech and Signal Processing . Toronto: IEEE, 2021: 1265-1269.
18 SUN W, SU Y, WU X, et al A novel end-to-end 1D-ResCNN model to remove artifact from EEG signals[J]. Neurocomputing, 2020, 404: 108- 121
doi: 10.1016/j.neucom.2020.04.029
19 HARTMANN K G, SCHIRRMEISTER R T, BALL T. EEG-GAN: generative adversarial networks for electroencephalograhic (EEG) brain signals [EB/OL]. (2018-06-05)[2023-10-01]. https://arxiv.org/pdf/1905.04149.
20 CORLEY I A, HUANG Y. Deep EEG super-resolution: upsampling EEG spatial resolution with generative adversarial networks [C]// IEEE EMBS International Conference on Biomedical and Health Informatics . Las Vegas: IEEE, 2018: 100-103.
21 LUO T, FAN Y, CHEN L, et al EEG signal reconstruction using a generative adversarial network with Wasserstein distance and temporal-spatial-frequency loss[J]. Frontiers in Neuroinformatics, 2020, 14: 15
doi: 10.3389/fninf.2020.00015
22 ANTONIADES A, SPYROU L, MARTIN-LOPEZ D, et al Deep neural architectures for mapping scalp to intracranial EEG[J]. International Journal of Neural Systems, 2018, 28 (8): 1850009
doi: 10.1142/S0129065718500090
23 HU M, CHEN J, JIANG S, et al E2SGAN: EEG-to-SEEG translation with generative adversarial networks[J]. Frontiers in Neuroscience, 2022, 16: 971829
doi: 10.3389/fnins.2022.971829
24 TSIOURIS Κ Μ, PEZOULAS V C, ZERVAKIS M, et al A long short-term memory deep learning network for the prediction of epileptic seizures using EEG signals[J]. Computers in Biology and Medicine, 2018, 99: 24- 37
doi: 10.1016/j.compbiomed.2018.05.019
25 EL-FIQI H, KASMARIK K, BEZERIANOS A, et al. Gate-layer autoencoders with application to incomplete EEG signal recovery [C]// International Joint Conference on Neural Networks . Budapest: IEEE, 2019: 1-8.
26 AL-MARRIDI A Z, MOHAMED A, ERBAD A. Convolutional auto-encoder approach for EEG compression and reconstruction in m-health systems [C]// 14th International Wireless Communications and Mobile Computing Conference . Chongqing: IEEE, 2018: 370-375.
27 JIAO Z, YOU H, YANG F, et al. Decoding EEG by visual guided deep neural networks [C]// International Joint Conferences on Artificial Intelligence . Macao: IEEE, 2019: 1387-1393.
28 YAO Y, PLESTED J, GEDEON T. Deep feature learning and visualization for EEG recording using autoencoders [C]// 25th International Conference on Neural Information Processing . Siem Reap: Springer, 2018: 554-566.
29 BASHIVAN P, RISH I, YEASIN M, et al. Learning representations from EEG with deep recurrent-convolutional neural networks [EB/OL]. (2016-02-29) [2023-10-01]. https://arxiv.org/pdf/1511.06448.
30 YU M, SUN Y, ZHU B, et al. Diverse frequency band-based convolutional neural networks for tonic cold pain assessment using EEG [J]. Neurocomputing , 2020, 378: 270-282.
31 PRASANTH T, THOMAS J, YUVARAJ R, et al. Deep learning for interictal epileptiform spike detection from scalp EEG frequency sub bands [C]// 42nd Annual International Conference of the IEEE Engineering in Medicine and Biology Society . Montréal: IEEE, 2020: 3703-3706.
32 SHEN F, PENG Y, KONG W, et al. Multi-scale frequency bands ensemble learning for EEG-based emotion recognition [EB/OL]. (2021-02-10)[2023-10-01].https://www.mdpi.com/1424-8220/21/4/1262.
33 MIAO M, ZHENG L, XU B, et al. A multiple frequency bands parallel spatial–temporal 3D deep residual learning framework for EEG-based emotion recognition [EB/OL]. (2022-09-18) [2023-10-01]. https://www.sciencedirect.com/science/article/abs/pii/S174680942200595X.
34 YAO Y, JP A, TG A. Information-preserving feature filter for short-term EEG signals [J]. Neurocomputing , 2020, 408: 91-99.
35 MIAO M, HU W, YIN H, et al. Spatial-frequency feature learning and classification of motor imagery EEG based on deep convolution neural network [EB/OL]. (2020-07-20)[2023-10-01]. https://www.hindawi.com/journals/cmmm/2020/1981728/.
36 LAWHERN V J, SOLON A J, WAYTOWICH N R, et al EE-GNet: a compact convolutional network for EEG-based brain-computer interfaces[J]. Journal of Neural Engineering, 2018, 15 (5): 056013.1- 056013.17
37 ZHAO D, TANG F, SI B, et al Learning joint space-time-frequency features for EEG decoding on small labeled data[J]. Neural Networks, 2019, 114: 67- 77
doi: 10.1016/j.neunet.2019.02.009
38 WANG J, LIANG S, HE D, et al. A sequential graph convolutional network with frequency-domain complex network of EEG signals for epilepsy detection [C]// IEEE International Conference on Bioinformatics and Bio-medicine . Seoul: IEEE, 2020.
39 ZHANG X, LU D, SHEN J, et al. Spatial-temporal joint optimization network on covariance manifolds of electroencephalography for fatigue detection [C]// IEEE International Conference on Bioinformatics and Biomedicine . Seoul: IEEE, 2020: 893-900.
40 FANG Z, WANG W, REN S, et al. Learning regional attention convolutional neural network for motion intention recognition based on EEG data [C]// 29th International Conference on International Joint Conferences on Artificial Intelligence . Yokohama: IEEE, 2021: 1570-1576.
41 ZHONG P, WANG D, MIAO C EEG-based emotion recognition using regularized graph neural networks[J]. IEEE Transactions on Affective Computing, 2020, 13 (3): 1290- 1301
42 SCHMIDT L A, TRAINOR L J Frontal brain electrical activity (EEG) distinguishes valence and intensity of musical emotions[J]. Cognition and Emotion, 2001, 15 (4): 487- 500
doi: 10.1080/02699930126048
43 LI Y, ZHENG W, ZONG Y, et al A bi-hemisphere domain adversarial neural network model for EEG emotion recognition[J]. IEEE Transactions on Affective Computing, 2018, 12 (2): 494- 504
44 SONG T, LIU S, ZHENG W, et al. Instance-adaptive graph for EEG emotion recognition [C]// AAAI Conference on Artificial Intelligence . New York: AAAI, 2020: 2701-2708.
45 LI R, WANG Y, LU B L. A multi-domain adaptive graph convolutional network for EEG-based emotion recognition [C]// 29th ACM International Conference on Multimedia . Chengdu: ACM, 2021: 5565-5573.
46 LI A, HUYNH C, FITZGERALD Z, et al Neural fragility as an EEG marker of the seizure onset zone[J]. Nature Neuroscience, 2021, 24 (10): 1465- 1474
doi: 10.1038/s41593-021-00901-w
47 GUNNARSDOTTIR K M, LI A, SMITH R J, et al Source-sink connectivity: a novel interictal EEG marker for seizure localization[J]. Brain, 2022, 145 (11): 3901- 3915
doi: 10.1093/brain/awac300
48 KOSTAS D, AROCA-OUELLETTE S, RUDZICZ F BENDR: using transformers and a contrastive self-supervised learning task to learn from massive amounts of EEG data[J]. Frontiers in Human Neuroscience, 2021, 15: 653659
doi: 10.3389/fnhum.2021.653659
49 DEVLIN J, CHANG M W, LEE K, et al. Bert: pre-training of deep bidirectional transformers for language understanding [EB/OL].(2019-05-24)[2023-10-01].https://arxiv.org/pdf/1810.04805.
50 JIA Z, LIM Y, WANG J, et al Multi-view spatial-temporal graph convolutional networks with domain generalization for sleep stage classification[J]. IEEE Transactions on Neural Systems and Rehabilitation Engineering, 2021, 29: 1977- 1986
doi: 10.1109/TNSRE.2021.3110665
51 LI Y, CHEN J, LI F, et al. Gmss: graph-based multi-task self-supervised learning for EEG emotion recognition [J]. IEEE Transactions on Affective Computing , 2022, 14(3): 2512-2525.
52 KUMAR V, REDDY L, SHARMA S K, et al. mulEEG: a multi-view representation learning on EEG signals [C]// Medical Image Computing and Computer Assisted Intervention. Singapore: Springer, 2022: 398-407.
53 MIKKELSEN K B, KIDMOSE P, HANSEN L K On the key hole hypothesis: high mutual information between ear and scalp EEG[J]. Frontiers in Human Neuroscience, 2017, 11 (341): 115
54 ANANDAKUMAR M, PRADEEPKUMAR J, KAPPEL S L, et al. A knowledge distillation framework for enhancing ear-EEG based sleep staging with scalp-EEG data [EB/OL]. (2022-10-27) [2023-10-01]. https://arxiv.org/pdf/2211.02638.
[1] 宦海,盛宇,顾晨曦. 基于遥感图像道路提取的全局指导多特征融合网络[J]. 浙江大学学报(工学版), 2024, 58(4): 696-707.
[2] 罗向龙,王亚飞,王彦博,王立新. 基于双向门控式宽度学习系统的监测数据结构变形预测[J]. 浙江大学学报(工学版), 2024, 58(4): 729-736.
[3] 宋明俊,严文,邓益昭,张俊然,涂海燕. 轻量化机器人抓取位姿实时检测算法[J]. 浙江大学学报(工学版), 2024, 58(3): 599-610.
[4] 钱庆杰,余军合,战洪飞,王瑞,胡健. 基于DL-BiGRU多特征融合的注塑件尺寸预测方法[J]. 浙江大学学报(工学版), 2024, 58(3): 646-654.
[5] 姚鑫骅,于涛,封森文,马梓健,栾丛丛,沈洪垚. 基于图神经网络的零件机加工特征识别方法[J]. 浙江大学学报(工学版), 2024, 58(2): 349-359.
[6] 孙雪菲,张瑞峰,关欣,李锵. 强化先验骨架结构的轻量型高效人体姿态估计[J]. 浙江大学学报(工学版), 2024, 58(1): 50-60.
[7] 郑超昊,尹志伟,曾钢锋,许月萍,周鹏,刘莉. 基于时空深度学习模型的数值降水预报后处理[J]. 浙江大学学报(工学版), 2023, 57(9): 1756-1765.
[8] 杨哲,葛洪伟,李婷. 特征融合与分发的多专家并行推荐算法框架[J]. 浙江大学学报(工学版), 2023, 57(7): 1317-1325.
[9] 李云红,段姣姣,苏雪平,张蕾涛,于惠康,刘杏瑞. 基于改进生成对抗网络的书法字生成算法[J]. 浙江大学学报(工学版), 2023, 57(7): 1326-1334.
[10] 权巍,蔡永青,王超,宋佳,孙鸿凯,李林轩. 基于3D-ResNet双流网络的VR病评估模型[J]. 浙江大学学报(工学版), 2023, 57(7): 1345-1353.
[11] 周欣磊,顾海挺,刘晶,许月萍,耿芳,王冲. 基于集成学习与深度学习的日供水量预测方法[J]. 浙江大学学报(工学版), 2023, 57(6): 1120-1127.
[12] 刘沛丰,钱璐,赵兴炜,陶波. 航空装配领域中命名实体识别的持续学习框架[J]. 浙江大学学报(工学版), 2023, 57(6): 1186-1194.
[13] 赵嘉墀,王天琪,曾丽芳,邵雪明. 基于GRU的扑翼非定常气动特性快速预测[J]. 浙江大学学报(工学版), 2023, 57(6): 1251-1256.
[14] 曹晓璐,卢富男,朱翔,翁立波,卢书芳,高飞. 基于草图的兼容性服装生成方法[J]. 浙江大学学报(工学版), 2023, 57(5): 939-947.
[15] 苏育挺,陆荣烜,张为. 基于注意力和自适应权重的车辆重识别算法[J]. 浙江大学学报(工学版), 2023, 57(4): 712-718.