基于图神经网络的多域特征融合表情识别算法研究

doi:10.3778/j.issn.1002-8331.2411-0297

摘要/Abstract

摘要： 针对现有深度模型需要大量训练数据且无法基于同一网络架构同时对宏表情和微表情进行识别的问题，提出一种基于空域与谱域特征级融合的图神经网络表情识别算法S2GNN（spatial and spectral graph neural networks）。以图像像素点作为节点，节点间的距离作为边，将面部建模为无向加权图，利用邻接矩阵存储图中的距离信息。得到邻接矩阵后，一方面基于图傅里叶变换将图像由空域转化为谱域，在谱域对图像进行卷积，得到图像的谱域特征。另一方面基于无向加权图在空域对图中节点及其相邻节点进行卷积，得到图像的空域特征。利用提出的基于空域特征引导的特征选择模块实现空域和谱域特征的融合。基于融合后的特征进行表情识别。实验结果表明，将空域特征和谱域特征相结合可以提高表情识别的准确率，即使在少量训练数据情况下也可取得较高的分类准确率。在8个公开数据集上进行测试，平均准确率可达89.11%。宏表情和微表情混合数据情况下，未加权F1值和未加权平均召回率分别为0.954和0.970。对实验结果进一步分析可以发现，眼部区域对提高遮挡情况下表情识别的准确率具有重要作用。

关键词: 图神经网络, 表情识别, 微表情识别, 空域特征引导

Abstract: To solve the problem that it is difficult to recognize macro-expressions and micro-expressions at the same time with a small amount of training data using deep learning, a multi-domain feature fusion framework based on graph neural networks is proposed, named S2GNN（spatial and spectral graph neural networks）. Firstly, pixels are represented by nodes and the distances between each node are represented by edges. The face can be built as an undirected weighted graph and the distances are saved as an adjacency matrix. On the one hand, images are transformed from spatial domain to spectral domain based on graph fourier transform, which can obtain spectral features. On the other hand, the node and its adjacent nodes are convoluted based on graph in spatial domain to obtain spatial domain features after adjacency matrix is obtained. And then, a spatial feature guided feature selection module is proposed to fuse the spatial and spectral domain features. Finally, results can be obtained based on the fused features. Results show that expressions recognition accuracy can be improved by combining spatial and spectral domain features and better performance can be obtained with a small amount of training data. An accuracy of 89.11% is reached based on eight open source datasets. In the case of mixed data with macro and micro expressions, the unweighted F1 score and unweighted average recall are 0.954 and 0.970, respectively. Furthermore, eye region plays an important role in improving facial expression recognition accuracy.

Key words: graph neural networks, facial expression recognition, micro expression recognition, spatial feature guided

张俊杰, 费程, 何伏刚. 基于图神经网络的多域特征融合表情识别算法研究[J]. 计算机工程与应用, 2025, 61(20): 182-193.

ZHANG Junjie, FEI Cheng, HE Fugang. Multi-Domain Feature Fusion for Facial Expression Recognition Based on Graph Neural Networks[J]. Computer Engineering and Applications, 2025, 61(20): 182-193.

参考文献

[1] ZENG H, SHU X, WANG Y, et al. EmotionCues: emotion-oriented visual summarization of classroom videos[J]. IEEE Transactions on Visulization and Computer Graphics, 2021, 27(7): 3168-3181.
[2] SAADI I, CUNNINGHAM D W, TALEB-AHMED A, et al. Driver’s facial expression recognition: a comprehensive survey[J]. Expert Systems with Applications, 2024, 242: 122784.
[3] ZHANG J J, ZHENG K, MAZHAR S, et al. Trusted emotion recognition based on multiple signals captured from video[J]. Expert Systems with Applications, 2023, 233: 120948.
[4] LI Y T, WEI J S, LIU Y, et al. Deep learning for micro-expression recognition: a survey[J]. IEEE Transactions on Affective Computing, 2022, 13(4): 2028-2046.
[5] LI Y, HUANG X, ZHAO G. Joint local and global information learning with single apex frame detection for micro-expression recognition[J]. IEEE Transactions on Image Processing, 2021, 30: 249-263.
[6] FANG Y C, LUO J, LOU C S. Fusion of multi-directional rotation invariant uniform LBP features for face recognition[C]//Proceedings of the 3rd International Symposium on Intelligent Information Technology Application. Piscataway: IEEE, 2009: 332-335.
[7] ZHANG T, ZHENG W M, CUI Z, et al. A deep neural network-driven feature learning method for multi-view facial expression recognition[J]. IEEE Transactions on Multimedia, 2016, 18(12): 2528-2536.
[8] KUMAR P, HAPPY S L, ROUTRAY A. A real-time robust facial expression recognition system using HOG features[C]//Proceedings of the International Conference on Computing, Analytics and Security Trends. Piscataway: IEEE, 2016: 289-293.
[9] SHOJAEILANGARI S, YAU W Y, NANDAKUMAR K, et al. Robust representation and recognition of facial emotions using extreme sparse learning[J]. IEEE Transactions on Image Processing, 2015, 24(7): 2140-2152.
[10] CHEN Y D, YANG X, CHAM T J, et al. Towards unbiased visual emotion recognition via causal intervention[C]//Proceedings of the 30th ACM International Conference on Multimedia. New York: ACM, 2022: 60-69.
[11] WANG L J, JIA G L, JIANG N, et al. EASE: robust facial expression recognition via emotion ambiguity-sensitive cooperative networks[C]//Proceedings of the 30th ACM International Conference on Multimedia. New York: ACM, 2022: 218-227.
[12] BARROS P, BARAKOVA E, WERMTER S. Adapting the interplay between personalized and generalized affect recognition based on an unsupervised neural framework[J]. IEEE Transactions on Affective Computing, 2022, 13(3): 1349-1365.
[13] KIM D H, BADDAR W J, JANG J, et al. Multi-objective based spatio-temporal feature representation learning robust to expression intensity variations for facial expression recognition[J]. IEEE Transactions on Affective Computing, 2019, 10(2): 223-236.
[14] VERMA M, VIPPARTHI S K, SINGH G, et al. LEARNet: dynamic imaging network for micro expression recognition[J]. IEEE Transactions on Image Processing, 2020, 29: 1618-1627.
[15] SONG B L, LI K, ZONG Y, et al. Recognizing spontaneous micro-expression using a three-stream convolutional neural network[J]. IEEE Access, 2019, 7: 184537-184551.
[16] XIA Z Q, HONG X P, GAO X Y, et al. Spatiotemporal recurrent convolutional networks for recognizing spontaneous micro-expressions[J]. IEEE Transactions on Multimedia, 2020, 22(3): 626-640.
[17] XIE Y, CHEN T S, PU T, et al. Adversarial graph representation adaptation for cross-domain facial expression recognition[C]//Proceedings of the 28th ACM International Conference on Multimedia. New York: ACM, 2020: 1255-1264.
[18] XU X, RUAN Z, YANG L. Facial expression recognition based on graph neural network[C]//Proceedings of the IEEE 5th International Conference on Image, Vision and Compu-ting. Piscataway: IEEE, 2020: 211-214.
[19] LIU D Z, ZHANG H T, ZHOU P. Video-based facial expression recognition using graph convolutional networks[C]//Proceedings of the 25th International Conference on Pattern Recognition. Piscataway: IEEE, 2021: 607-614.
[20] ZHOU J Z, ZHANG X M, LIU Y, et al. Facial expression recognition using spatial-temporal semantic graph network[C]//Proceedings of the 2020 IEEE International Conference on Image Processing. Piscataway: IEEE, 2020: 1961-1965.
[21] 黄苗苗, 王慧颖, 王梅霞, 等. 图嵌入学习研究综述: 从简单图到复杂图[J/OL]. 计算机科学, 1-29[2025-10-11]. https://link.cnki.net/urlid/50.1075.tp.20250613.1315.030.
HUANG M M，WANG H Y, WANG M X, et al.Survey of graph embedding learning: from simple graphs to complex graphs[J/OL]. Computer Science, 1-29[2025-10-11]. https://link.cnki.net/urlid/50.1075.tp.20250613.1315.030.
[22] 周诚辰, 于千城, 张丽丝, 等. Graph Transformers研究进展综述[J]. 计算机工程与应用, 2024, 60(14): 37-49.
ZHOU C C, YU Q C, ZHANG L S, et al. Overview of rese-arch progress in graph Transformers[J]. Computer Engine-ering and Applications, 2024, 60(14): 37-49.
[23] GORI M, MONFARDINI G, SCARSELLI F. A new model for learning in graph domains[C]//Proceedings of the IEEE International Joint Conference on Neural Networks. Piscataway: IEEE, 2005: 729-734.
[24] MICHELI A. Neural network for graphs: a contextual constructive approach[J]. IEEE Transactions on Neural Networks, 2009, 20(3): 498-511.
[25] BRUNA J, ZAREMBA W, SZLAM A, et al. Spectral networks and locally connected networks on graphs[J]. arXiv: 1312.62031, 2013.
[26] KIPF T, WELLING M. Semi-supervised classification with graph convolutional networks[J]. arXiv:1609.02907, 2016.
[27] EKMAN P, FRIESEN W V. Measuring facial movement[J]. Environmental Psychology and Nonverbal Behavior, 1976, 1(1): 56-75.
[28] ZHANG J J, FEI C, ZHENG Y Q, et al. Trusted emotion recognition based on multiple signals captured from video and its application in intelligent education[J]. Electronic Research Archive, 2024, 32(5): 3477-3521.
[29] LUCEY P, COHN J F, KANADE T, et al. The extended Cohn-Kanade dataset (CK+): a complete dataset for action unit and emotion-specified expression[C]//Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE, 2010: 94-101.
[30] LYONS M, AKAMATSU S, KAMACHI M, et al. Coding facial expressions with Gabor wavelets[C]//Proceedings of the 3rd IEEE International Conference on Automatic Face and Gesture Recognition. Piscataway: IEEE, 1998: 200-205.
[31] CHEN L F, YEN Y S. Taiwanese facial expression image database[EB/OL]. (2020-5-12) [2024-08-19]. http://bml.ym.edu.tw/tfeid/.
[32] LI S, DENG W H. Reliable crowdsourcing and deep locality-preserving learning for unconstrained facial expression recognition[J]. IEEE Transactions on Image Processing, 2019, 28(1): 356-370.
[33] YAN W J, WU Q, LIU Y J, et al. CASME database: a dataset of spontaneous micro-expressions collected from neutralized faces[C]//Proceedings of the 10th IEEE International Conference and Workshops on Automatic Face and Gesture Recognition. Piscataway: IEEE, 2013: 1-7.
[34] YAN W J, LI X, WANG S J, et al. CASME II: an improved spontaneous micro-expression database and the baseline evaluation[J]. PLoS One, 2014, 9(1): e86041.
[35] LI X B, PFISTER T, HUANG X H, et al. A spontaneous micro-expression database: inducement, collection and baseline[C]//Proceedings of the 10th IEEE International Conference and Workshops on Automatic Face and Gesture Recognition. Piscataway: IEEE, 2013: 1-6.
[36] DAVISON A K, LANSLEY C, COSTEN N, et al. SAMM: a spontaneous micro-facial movement dataset[J]. IEEE Transactions on Affective Computing, 2018, 9(1): 116-129.
[37] PFISTER T, LI X B, ZHAO G Y, et al. Differentiating spontaneous from posed facial expressions within a generic facial expression recognition framework[C]//Proceedings of the IEEE International Conference on Computer Vision Workshops. Piscataway: IEEE, 2011: 868-875.
[38] WANG F P, LI J, QI C, et al. JGULF: joint global and unilateral local feature network for micro-expression recogn-ition[J]. Image and Vision Computing, 2024, 147: 105091.
[39] LIONG S T, GAN Y S, SEE J, et al. Shallow triple stream three-dimensional CNN (STSTNet) for micro-expression recognition[C]//Proceedings of the 14th IEEE International Conference on Automatic Face & Gesture Recognition. Piscataway: IEEE, 2019: 1-5.
[40] GAN Y S, LIONG S T, YAU W C, et al. OFF-ApexNet on micro-expression recognition system[J]. Signal Processing: Image Communication, 2019, 74: 129-139.
[41] KRIZHEVSKY A, SUTSKEVER I, HINTON G E. ImageNet classification with deep convolutional neural networks[J]. Communications of the ACM, 2017, 60(6): 84-90.
[42] IANDOLA F N, HAN S, MOSKEWICZ M W, et al. Squee-zeNet: AlexNet-level accuracy with 50x fewer parameters and <0.5MB model size[J]. arXiv：1602.07360, 2016.
[43] 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. Piscataway: IEEE, 2015: 1-9.
[44] SIMONYAN K, ZISSERMAN A. Very deep convolutional networks for large-scale image recognition [J]. arXiv:1409. 1556, 2014.
[45] YANG J J, WANG S F. Capturing spatial and temporal patterns for distinguishing between posed and spontaneous expressions[C]//Proceedings of the 25th ACM International Conference on Multimedia. New York: ACM, 2017: 469-477.
[46] WANG S F, HAO L F, JI Q. Posed and spontaneous expression distinction using latent regression Bayesian networks[J]. ACM Transactions on Multimedia Computing, Communications, and Applications, 2020, 16(3): 1-18.
[47] WANG S F, WU C L, HE M H, et al. Posed and spontaneous expression recognition through modeling their spatial patterns[J]. Machine Vision and Applications, 2015, 26(2): 219-231.
[48] WANG S F, WU C L, JI Q. Capturing global spatial patterns for distinguishing posed and spontaneous expressions[J]. Computer Vision and Image Understanding, 2016, 147: 69-76.
[49] KARTHEEK M N, PRASAD M V N K, BHUKYA R. DRCP: dimensionality reduced chess pattern for person independent facial expression recognition[J]. International Journal of Pattern Recognition and Artificial Intelligence, 2022, 36(11): 2256016.
[50] BAYGIN M, TUNCER I, DOGAN S, et al. Automated facial expression recognition using exemplar hybrid deep feature generation technique[J]. Soft Computing, 2023, 27(13): 8721-8737.
[51] CHIRRA V R R, UYYALA S R, KOLLI V K K. Virtual facial expression recognition using deep CNN with ensemble learning[J]. Journal of Ambient Intelligence and Humanized Computing, 2021, 12(12): 10581-10599.
[52] WANG S F, ZHENG Z Q, YIN S, et al. A novel dynamic model capturing spatial and temporal patterns for facial expression analysis[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2020, 42(9): 2082-2095.
[53] MOEINI A, FAEZ K, SADEGHI H, et al. 2D facial expression recognition via 3D reconstruction and feature fusion[J]. Journal of Visual Communication and Image Representation, 2016, 35: 1-14.
[54] MEHTA A, RASTEGARI M. Separable self-attention for mobile vision transformers[J]. arXiv:2206.02680, 2022.
[55] MAAZ M, SHAKER A, CHOLAKKAL H, et al. EdgeNeXt: efficiently amalgamated CNN-transformer architecture for mobile vision applications[C]//Proceedings of the European Conference on Computer Vision. Cham: Springer Nature Switzerland, 2023: 3-20.
[56] CHEN J R, KAO S H, 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. Piscataway: IEEE, 2023: 12021-12031.
[57] SUN M M, YAN C M. FGENet: a lightweight facial expression recognition algorithm based on FasterNet[J]. Signal, Image and Video Processing, 2024, 18(8): 5939-5956.
[58] BORING E G. Titchener’s experimentalists[J]. Journal of the History of the Behavioral Sciences, 1967, 3(4): 315-325.
[59] HUGHES M A. Emotions revealed: recognizing faces and feelings to improve communication and emotional life[J]. Lib-rary Journal, 2003, 128(8): 140.
[60] IZARD C, FINE S, SCHULTZ D, et al. Emotion knowledge as a predictor of social behavior and academic competence in children at risk[J]. Psychological Science, 2001, 12(1): 18-23.