基于深度残差网络的人脸识别算法研究综述
A Review of Research on Face Recognition Algorithms Based on Deep Residual Networks
DOI: 10.12677/jisp.2026.152021, PDF,    科研立项经费支持
作者: 方小红, 王小生, 刘超飞*：江西理工大学理学院，江西 赣州；郭桥生：朝阳聚声泰(信丰)科技有限公司，江西 赣州
关键词: 深度残差网络；人脸识别；恒等变换；卷积神经网络；计算机视觉；Deep Residual Networks； Face Recognition； Identity Transformation； Convolutional Neural Networks； Computer Vision
摘要: 深度学习在各种各样的问题上都取得了非常好的表现，比如物体识别、语音识别和自然语音处理。在不同类型的深度神经网络中，深度残差网络(ResNet)得到了广泛的研究，为人脸识别提供了新的解决方案。因此，本文从深度残差网络的设计出发，介绍了深度残差网络的演进及变体，以及在人脸识别领域中常用的数据集及优化器选择、正则化方法，之后介绍了深度残差网络及改进，包括残差块重构、引入注意力机制、轻量化ResNet、构建ResNet系统，在人脸识别任务中的表现。
Abstract: Deep learning has achieved remarkable performance across a wide variety of tasks, such as object recognition, speech recognition, and natural language processing. Among different types of deep neural networks, Deep Residual Networks (ResNet) have been extensively studied, providing new solutions for face recognition. Therefore, starting from the design of Deep Residual Networks, this paper introduces their evolution and variants, as well as commonly used datasets, optimizer selection, and regularization methods in the field of face recognition. Subsequently, it discusses the performance of Deep Residual Networks and their improvements in face recognition tasks, such as residual block reconstruction, the incorporation of attention mechanisms, lightweight ResNet designs, and the construction of ResNet systems.
文章引用：方小红, 王小生, 刘超飞, 郭桥生. 基于深度残差网络的人脸识别算法研究综述[J]. 图像与信号处理, 2026, 15(2): 248-263. https://doi.org/10.12677/jisp.2026.152021

参考文献

[1] Gu, J., Wang, Z., Kuen, J., Ma, L., Shahroudy, A., Shuai, B., et al. (2018) Recent Advances in Convolutional Neural Networks. Pattern Recognition, 77, 354-377. [Google Scholar] [CrossRef] 
[2] Alshehri, A. (2024) Skin-NeT: Skin Cancer Diagnosis Using VGG and ResNet-Based Ensemble Learning Approaches. Traitement du Signal, 41, 1689-1705. [Google Scholar] [CrossRef] 
[3] Tufenkci, S. and Alagoz, B.B. (2026) Improved Classification of Star and Galaxy from Telescope by Using a Spatio-Spectral Feature ResNet Model. Advances in Space Research, 77, 5425-5439. [Google Scholar] [CrossRef] 
[4] Rubin, J., Parvaneh, S., Rahman, A., Conroy, B. and Babaeizadeh, S. (2018) Densely Connected Convolutional Networks for Detection of Atrial Fibrillation from Short Single-Lead ECG Recordings. Journal of Electrocardiology, 51, S18-S21. [Google Scholar] [CrossRef] [PubMed]
[5] LeCun, Y., Boser, B., Denker, J.S., Henderson, D., Howard, R.E., Hubbard, W., et al. (1989) Backpropagation Applied to Handwritten Zip Code Recognition. Neural Computation, 1, 541-551. [Google Scholar] [CrossRef] 
[6] Krizhevsky, A., Sutskever, I. and Hinton, G.E. (2012) ImageNet Classifification with Deep CNNs. Proceedings of Conference and Workshop on Neural Information Processing Systems (NIPS) 2012, Lake Tahoe, 3-6 December 2012, 1097-1105.
[7] He, K., Zhang, X., Ren, S. and Sun, J. (2016) Deep Residual Learning for Image Recognition. 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Las Vegas, 27-30 June 2016, 770-778. [Google Scholar] [CrossRef] 
[8] Liu, Q. and Bai, Z. (2026) 18F-FDG PET/CT Imaging Features-Driven Construction of Optimized CNN Model: For Post-Exercise Lower Limb Muscle Metabolism Assessment in Sanda Athletes. Journal of Radiation Research and Applied Sciences, 19, Article ID: 102215. [Google Scholar] [CrossRef] 
[9] Nguyen, Q.D. and Thai, H. (2026) A Robust and Efficient CNN-Transformer Network for Crack Segmentation of High Resolution Images. Engineering Structures, 353, Article ID: 122159. [Google Scholar] [CrossRef] 
[10] Shi, Q., Zhang, Q., Xi, X. and Long, L. (2026) Study on Dual Early Warning of Coal Spontaneous Combustion Based on the BO-CNN-GRU Model. Process Safety and Environmental Protection, 208, Article ID: 108535. [Google Scholar] [CrossRef] 
[11] Afzal, I., Jichang, G., Siddiqui, F. and Fahad, M. (2026) CNN-CECA: Underwater Image Enhancement via CNN-Driven Nonlinear Curve Estimation and Channel-Wise Attention in Multi-Color Spaces. Image and Vision Computing, 167, Article ID: 105916. [Google Scholar] [CrossRef] 
[12] Zhao, W., Xing, R., Wei, C., Sun, B., Jiang, T. and Zhao, Q. (2026) Improved Boundary-Aware Mask R-CNN Using Stereo Vision for Automated Rebar Inspection. Automation in Construction, 183, Article ID: 106801. [Google Scholar] [CrossRef] 
[13] Liu, X., Wang, M., Yang, P., Wang, W., He, L. and Zhu, H. (2026) A GASF-ResNet18 Method for the Calibration of Hydrophones with Optical Technique in a Non-Free-Field Water Tank. Measurement, 266, Article ID: 120404. [Google Scholar] [CrossRef] 
[14] Zhang, Y., Zhan, Y., Chen, H., Zhu, Y., Zhao, L. and Tian, Y. (2026) Research on Transmission Line Icing Classification and Recognition Algorithm Based on BiTex-ResNet34. Cold Regions Science and Technology, 245, Article ID: 104845. [Google Scholar] [CrossRef] 
[15] Salama, W.M. and Aly, M.H. (2026) Brain Tumor Segmentation and Classification: A CVAE-UNETR-ResNet50-VGG16 Hybrid Deep Learning Approach. Alexandria Engineering Journal, 135, 433-449. [Google Scholar] [CrossRef] 
[16] Mahmood Khan, R.N., Hussain, L., Alluhaidan, A.S., Majid, A., Lone, K.J., Verdiyev, R., et al. (2025) COVID-19 Lung Infection Detection Using Deep Learning with Transfer Learning and ResNet101 Features Extraction and Selection. Waves in Random and Complex Media, 35, 7886-7909. [Google Scholar] [CrossRef] 
[17] K, S., Nikhat, H.F., Poonkodi, S. and R, P. (2024) A Novel Approach for Lung Cancer Segmentation and Classification Using MSF-Customized ResNet152. IETE Journal of Research, 71, 1050-1062. [Google Scholar] [CrossRef] 
[18] Cao, Q., Shen, L., Xie, W., Parkhi, O.M. and Zisserman, A. (2018) VGGFace2: A Dataset for Recognising Faces across Pose and Age. 2018 13th IEEE International Conference on Automatic Face & Gesture Recognition (FG 2018), Xi’an, 15-19 May 2018, 67-74. [Google Scholar] [CrossRef] 
[19] Huang, G.B., Mattar, M., Berg, T., et al. (2008) Labeled Faces in the Wild: A Database ForStudying Face Recognition in Unconstrained Environments. Workshop on Faces in ‘Real-Life’ Images: Detection, Alignment, and Recognition 2008, Marseille, October 2008, 1-14.
[20] Erakin, M.E., Demir, U. and Ekenel, H.K. (2021) On Recognizing Occluded Faces in the Wild. 2021 International Conference of the Biometrics Special Interest Group (BIOSIG), Darmstadt, 15-17 September 2021, 1-5. [Google Scholar] [CrossRef] 
[21] Rodrigo, M., Gonzalez-Sosa, E., Cuevas, C. and Garcia, N. (2022) UPM-GTI-Face: A Dataset for the Evaluation of the Impact of Distance and Masks in Face Detection and Recognition Systems. 2022 18th IEEE International Conference on Advanced Video and Signal Based Surveillance (AVSS), Madrid, 29 November-2 December 2022, 1-8. [Google Scholar] [CrossRef] 
[22] Grgic, M., Delac, K. and Grgic, S. (2011) SCface—Surveillance Cameras Face Database. Multimedia Tools and Applications, 51, 863-879. [Google Scholar] [CrossRef] 
[23] Liliana, D.Y. (2019) Emotion Recognition from Facial Expression Using Deep Convolutional Neural Network. Journal of Physics: Conference Series, 1193, Article ID: 012004. [Google Scholar] [CrossRef] 
[24] Yamazaki, M., Kasagi, A., Tabuchi, A., et al. (2019) Yet Another Accelerated SGD: ResNet-50 Training on ImageNet in 74.7 Seconds. arXiv: 1903.12650.
[25] Sudewo, E.B., Kunta Biddinika, M., Umar, R. and Fadlil, A. (2025) Evaluating the Impact of Optimizer Hyperparameters on ResNet in Hanacaraka Character Recognition. Preservation, Digital Technology & Culture, 54, 113-123. [Google Scholar] [CrossRef] 
[26] Rumelhart, D.E., Hinton, G.E. and Williams, R.J. (1986) Learning Representations by Back-Propagating Errors. Nature, 323, 533-536. [Google Scholar] [CrossRef] 
[27] Azida Muhammad, N., Ab Nasir, A., Ibrahim, Z. and Sabri, N. (2018) Evaluation of CNN, AlexNet and GoogleNet for Fruit Recognition. Indonesian Journal of Electrical Engineering and Computer Science, 12, 468-475. [Google Scholar] [CrossRef] 
[28] Praveen, T.N.V.S., Sivathmika, D., Jahnavi, G. and Bolledu, J. (2023) An In-Depth Exploration of ResNet-50 for Complex Emotion Recognition to Unraveling Emotional States. 2023 International Conference on Advancement in Computation & Computer Technologies (InCACCT), Gharuan, 5-6 May 2023, 1-5. [Google Scholar] [CrossRef] 
[29] 金力, 金正贤, 卢海妹, 等. 基于ResNet算法的课堂教学效果评价模型[J]. 兰州文理学院学报(自然科学版), 2021, 35(6): 62-66.
[30] Liu, J. (2024) Face Recognition Technology Based on ResNet-50. Applied and Computational Engineering, 39, 160-165. [Google Scholar] [CrossRef] 
[31] 易锋, 胡馨莹. 基于深度残差网络的行人人脸识别算法研究[J]. 电脑知识与技术, 2018, 14(23): 233-235.
[32] 卢官明, 朱海锐, 郝强, 等. 基于深度残差网络的人脸表情识别[J]. 数据采集与处理, 2019, 34(1): 50-57.
[33] Jain, A., Moparthi, N.R., Swathi, A., Sharma, Y.K., Mittal, N., Alhussen, A., et al. (2024) Deep Learning-Based Mask Identification System Using ResNet Transfer Learning Architecture. Computer Systems Science and Engineering, 48, 341-362. [Google Scholar] [CrossRef] 
[34] Wangean, D.A., Pangestu, G., Setyawan, S., Maulana, F.I., Gunawan, E.P. and Huda, C. (2024) The Implementation of ResNet-50 Architecture for Face Recognition Algorithm in Attendance System. AIP Conference Proceedings, 2927, Article ID: 060014. [Google Scholar] [CrossRef] 
[35] Kumar, M., Gulhane, M., Kumar, S., Sharma, H., Verma, R. and Verma, D. (2023) Improved Multi-Face Detection with ResNet for Real-World Applications. 2023 12th International Conference on System Modeling & Advancement in Research Trends (SMART), Moradabad, 22-23 December 2023, 43-49. [Google Scholar] [CrossRef] 
[36] Medjdoubi, A., Meddeber, M. and Yahyaoui, K. (2024) Smart City Surveillance: Edge Technology Face Recognition Robot Deep Learning Based. International Journal of Engineering, 37, 25-36. [Google Scholar] [CrossRef] 
[37] 方国康, 李俊, 王垚儒. 基于深度学习的ARM平台实时人脸识别[J]. 计算机应用, 2019, 39(8): 2217-2222.
[38] 张立军, 曹江涛, 姬晓飞, 等. 基于深度学习的学生课堂学习状态监测系统设计[J]. 电子测量与仪器学报, 2024, 38(4): 37-45.
[39] Rodrigo, M., Cuevas, C. and García, N. (2024) Comprehensive Comparison between Vision Transformers and Convolutional Neural Networks for Face Recognition Tasks. Scientific Reports, 14, Article No. 21392. [Google Scholar] [CrossRef] [PubMed]
[40] Komlavi, A.A., Chaibou, K. and Naroua, H. (2024) Comparative Study of Machine Learning Algorithms for Face Recognition. Revue Africaine de Recherche En Informatique et Mathématiques Appliquées, 40, 1-26.
[41] Nguyen, T.V. and Chu, T.D. (2023) Comparative Study on the Performance of Face Recognition Algorithms. Eureka: Physics and Engineering, No. 4, 120-132. [Google Scholar] [CrossRef] 
[42] Safwat, S., Mahmoud, A., Eldesouky Fattoh, I. and Ali, F. (2024) Hybrid Deep Learning Model Based on GAN and RESNET for Detecting Fake Faces. IEEE Access, 12, 86391-86402. [Google Scholar] [CrossRef] 
[43] 陈希彤, 卢涛. 基于全局深度分离卷积残差网络的高效人脸识别算法[J]. 武汉工程大学学报, 2019, 41(3): 276-282.
[44] 周婕, 马明栋. 基于改进的ResNet网络的人脸表情识别[J]. 计算机技术与发展, 2022, 32(1): 25-29.
[45] Deng, Z., Chiang, H., Kang, L. and Li, H. (2023) A Lightweight Deep Learning Model for Real-Time Face Recognition. IET Image Processing, 17, 3869-3883. [Google Scholar] [CrossRef] 
[46] 倪锦园, 张建勋, 张馨月. 深度宽残差网络注意力机制的人脸表情识别[J]. 重庆理工大学学报(自然科学), 2023, 37(1): 177-185.
[47] 张栋昱, 赵磊. 融合注意力机制改进ResNet的人脸表情识别[J]. 计算机技术与发展, 2023, 33(5): 130-137.
[48] 石敏. 基于深度残差网络与人脸关键点的表情识别[J]. 软件导刊, 2022, 21(4): 226-232.
[49] Lai, Z., Chen, R., Jia, J. and Qian, Y. (2023) Real-Time Micro-Expression Recognition Based on ResNet and Atrous Convolutions. Journal of Ambient Intelligence and Humanized Computing, 14, 15215-15226. [Google Scholar] [CrossRef] 
[50] 高涛, 邵倩, 张亚南, 等. 基于深度残差网络的人脸表情识别研究[J]. 电子设计工程, 2020, 28(23): 101-104.
[51] 刘瑞明, 徐春融, 周韬, 等. 基于深度残差网络模型压缩的局部遮挡人脸识别[J]. 智能计算机与应用, 2023, 13(11): 244-249.
[52] 贺怀清, 闫建青, 惠康华. 基于深度残差网络的轻量级人脸识别方法[J]. 计算机应用, 2022, 42(7): 2030-2036.
[53] 吴宇豪, 陈晓辉. 基于改进的ResNet的人脸表情识别系统[J]. 信息通信, 2020(7): 37-39.
[54] Obaid, A.M., Shawkat, A.S. and Abdulhussein, N.S. (2023) Exploring the Potential of A-ResNet in Person-Independent Face Recognition and Classification. International Journal of Advanced Network, Monitoring and Controls, 8, 12-19. [Google Scholar] [CrossRef] 
[55] 苗壮, 程卫月, 林克正, 等. 融合并行网络特征的人脸表情识别算法[J]. 哈尔滨理工大学学报, 2022, 27(6): 95-102.
[56] 李靖宇, 程卫月, 李子翔, 等. 超分辨率重建的微小人脸识别算法[J]. 哈尔滨理工大学学报, 2022, 27(3): 52-58.
[57] 刘静. 基于深度残差网络的光照干扰人脸识别方法[J]. 通化师范学院学报, 2024, 45(10): 42-47.
[58] Dai, C. and Zeng, X. (2024) Occluded Face Recognition Network Based on DCGAN and ResNet. Procedia Computer Science, 243, 724-733. [Google Scholar] [CrossRef] 
[59] Istiqomah, A.A., Sari, C.A., Susanto, A. and Rachmawanto, E.H. (2024) Facial Expression Recognition Using Convolutional Neural Networks with Transfer Learning ResNet-50. Journal of Applied Informatics and Computing, 8, 257-264. [Google Scholar] [CrossRef] 
[60] Sunil, M.P. and Hariprasad, S.A. (2023) Facial Emotion Recognition Using a Modified Deep Convolutional Neural Network Based on the Concatenation of XCEPTION and RESNET50 V2. International Journal of Electrical and Electronics Engineering, 10, 94-105. [Google Scholar] [CrossRef] 
[61] 马玉环, 张瑞军, 武晨, 等. 深度残差网络和LSTM结合的图像序列表情识别[J]. 重庆邮电大学学报(自然科学版), 2020, 32(5): 874-883.
[62] 黄健. 基于深度学习与二维离散小波分解特征相融合的adaboost人脸识别模型[J]. 软件工程, 2020, 23(2): 43-46.
[63] Dwivedi, R.K. and Kumar, D. (2023) ResNet and PCA-Based Deep Learning Scheme for Efficient Face Recognition. International Journal of Intelligent Information Technologies, 19, 1-20. [Google Scholar] [CrossRef] 
[64] Koodalsamy, B., Veerayan, M.B. and Narayanasamy, V. (2023) Face Recognition Using Deep Learning. E3S Web of Conferences, 387, Article ID: 05001. [Google Scholar] [CrossRef] 
[65] 黄奥云, 程宾洋, 孙家炜. 基于卷积内部进化机制与特征融合的人脸识别算法的研究[J]. 现代计算机(专业版), 2019, 25(10): 75-80.
[66] Saravanan, P., Pandey, A., Joshi, K., Rondon, R., Narasimharao, J. and Imran, A.A. (2023) Using Machine Learning Principles, the Classification Method for Face Spoof Detection in Artificial Neural Networks. 2023 3rd International Conference on Advance Computing and Innovative Technologies in Engineering (ICACITE), Greater Noida, 12-13 May 2023, 2784-2788. [Google Scholar] [CrossRef] 
[67] Thanathamathee, P., Sawangarreerak, S., Kongkla, P. and Mohd Nizam, D.N. (2023) An Optimized Machine Learning and Deep Learning Framework for Facial and Masked Facial Recognition. Emerging Science Journal, 7, 1173-1187. [Google Scholar] [CrossRef]