Robust Micro-expression Recognition Based on Fused Optimized VGG16 Architecture

DOI: https://doi.org/10.62517/jike.202504406

Author(s)

Senlin Zhang1, Shuang Liang1, Junhao Shi1, Xiaofeng Li2,*

Affiliation(s)

1College of Artificial Intelligence and Big Data, Henan University of Technology, Zhengzhou, China 2Anhui USTC iFLYTEK Co., Ltd., Anhui, Hefei, China * Corresponding Author

Abstract

Due to the short duration and subtle changes of facial expressions, traditional methods for micro expression recognition struggle to balance high accuracy and real-time performance. In response to this issue, this article proposes a robust recognition method based on fusion optimized VGG16 architecture, focusing on addressing the limitations of traditional VGG16 models such as large parameter count, low training efficiency, and insufficient capture of subtle features. By introducing depthwise separable convolution to reduce computational complexity, combined with CBAM attention mechanism to enhance the ability to focus on key region features, PReLU activation function is used to optimize nonlinear feature expression, and residual connection structure is designed to alleviate gradient vanishing problem. Experiments on the CK+and CASME2 datasets showed that the improved model achieved an accuracy of 94.90% on the CK+dataset (with only 1.77% of the original model parameters), and an accuracy of 98.23% on the CASME2 dataset, datasets showed that the improved model achieved an accuracy of 94.90% on the CK + dataset (with only 1.77% of the dataset), significantly better than the traditional VGG16 model. The ablation experiment verified the effectiveness of each module and provided an effective and feasible solution for real-time micro expression detection.

Keywords

Micro Expression Recognition; VGG16; Residual Connection; Attention Mechanism; Depthwise Separable Convolution

References

[1]Zhu C, Chen X, Zhang J, et al. Comparison of Ecological Micro-Expression Recognition in Patients with Depression and Healthy Individuals. Frontiers in Behavioral Neuroscience, 2017. [2]Islam S, Saha P, Chowdhury T, et al. Non-invasive Deception Detection in Videos Using Machine Learning Techniques//2021 5th International Conference on Electrical Engineering and Information Communication Technology (ICEEICT), Dhaka, Bangladesh. 2021: 1-6. [3]K Pe, ANJAN S, T Nn, et al. Enhancing Human-Computer Interaction through Emotion Recognition in Real-Life Speech. 2023. [4]Zeng Yi, Wang Guoqiang, Jiang Dongchen. Micro facial expression recognition method based on multi-scale ShuffleNet. Journal of natural science of heilongjiang university, 2024, 9 (6) : 718-730. The DOI: 10.13482 / j.i ssn1001-7011.2024.11.072. [5]Shreve M, Godavarthy S, Manohar V, et al. Towards macro- and micro-expression spotting in video using strain patterns//2009 Workshop on Applications of Computer Vision (WACV), Snowbird, UT, USA. 2009: 1-6. [6]Hui T W, Tang X, Loy C C. LiteFlowNet: A Lightweight Convolutional Neural Network for Optical Flow Estimation//2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition, Salt Lake City, UT, USA. 2018. [7]He K, Zhang X, Ren S, et al. Delving Deep into Rectifiers: Surpassing Human-Level Performance on ImageNet Classification//2015 IEEE International Conference on Computer Vision (ICCV). Santiago, Chile: IEEE, 2015: 1026-1034. [8]He K, Zhang X, Ren S, et al. Deep Residual Learning for Image Recognition//2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Las Vegas, NV, USA. 2016. [9]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//2010 IEEE Computer Society Conference on Computer Vision and Pattern Recognition - Workshops, San Francisco, CA, USA. 2010. [10]Qu F, Wang S J, Yan W J, et al. CAS(ME): A Database for Spontaneous Macro-Expression and Micro-Expression Spotting and Recognition. IEEE Transactions on Affective Computing, 2018: 424-436. [11]Lin T Y, Goyal P, Girshick R, et al. Focal Loss for Dense Object Detection//2017 IEEE International Conference on Computer Vision (ICCV). Venice, Italy: IEEE, 2017: 2999-3007. [12]HE K, ZHANG X, REN S, et al. Deep Residual Learning for Image Recognition//2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Las Vegas, NV, USA. 2016. [13]TAN M, LE QuocV. EfficientNet: Rethinking Model Scaling for Convolutional Neural Networks. 2019. [14]Ma N, Zhang X, Zheng H T, et al. ShuffleNet V2: Practical Guidelines for Efficient CNN Architecture Design //European Conference on Computer Vision (ECCV). Munich, Germany: Springer, 2018: 122-138. [15]Howard A, Sandler M, Chu G, et al. Searching for MobileNetV3//2019 IEEE/CVF International Conference on Computer Vision (ICCV). Seoul, Korea (South): IEEE, 2019: 1314-1324.