Article Details

Title · Authors · Abstract · Keywords · DOI · References

Research on Generation and Recognition of Few-Shot Miao Embroidery Images

WU Jing1,2, YANG Bangqin1, ZHANG Yinjian1, LI Mingzhu2, CHEN Yan1,3

(1.School of Data Science and Information Engineering, Guizhou Minzu University, Guiyang 550025, China; 2.ZX-YZ School of Network Science, Haikou University of Economics, Haikou 570203, China; 3.School of Automation Science and Engineering, South China University of Technology, Guangzhou 510641, China)

Section：Computer Technology · Vol. 2025, Issue 2 · Pages：24-32 · 2096-4706（2025）02-0024-09 · CLC number：TP391 · Document code：A · DOI：10.19850/j.cnki.2096-4706.2025.02.005

Abstract

To address the issues of insufficient sample sizes and low recognition accuracy of Miao embroidery images, this paper proposes a novel Miao embroidery image classification model, which combines StyleGAN2 with optimized ResNet50 by the Efficient Channel Attention (ECA). Firstly, StyleGAN2 is utilized to train and generate on the original Miao embroidery dataset, thereby augmenting the Few-Shot dataset. Subsequently, the ECA Attention Mechanism is integrated into the ResNet50 backbone network to enhance the feature extraction capability of the model. Ultimately, the new StyleGAN2-ECA-ResNet50 model is created by combining StyleGAN2 and ResNet50 optimized by ECA for the recognition of Few-Shot Miao embroidery images. Experimental results show that the accuracy of this method reaches 89.29% on the test set, which is an improvement of 5.87% over the traditional ResNet50 model and surpasses several advanced Deep Learning models in performance.

Keywords

Miao embroidery; Few-Shot image classification; data augmentation; StyleGAN2; ECA

References

  1. [1] Geng Xin entered Comparative Analysis of Colors in Guizhou Miao Embroidery and Fauvism Painting [J]. Literary Competition, 2024 (7): 194-197
  2. [2] Liu Wenliang, Huang Jie Research on the confusion and solution path of intelligent inheritance of Miao embroidery in the digital age [J]. Furniture and Interior Decoration, 2023, 30 (1): 68-74
  3. [3] Dai Yongqi, Peng Li, Xie Naipeng Digital Empowerment: Digital Extraction of Intangible Cultural Heritage Miao Embroidery Patterns Based on Visual Transformer [J]. Silk, 2024, 61 (7): 14-24
  4. [4] Wang Jianping, Zhou Yuying, Yu Qianwen, etc Research on the Development and Process Parameters of Digital Simulation Crepe Embroidery Based on Intelligent Embroidery Machine [J]. Silk, 2024, 61 (7): 47-54
  5. [5] Chen Shijie, Wang Weixing, Peng Li Research on Pattern Segmentation Algorithm of Miao Embroidery Pieces Based on Multi scale Network [J]. Computer Technology and Development, 2023, 33 (11): 149-155
  6. [6] Tian Yujing, Xue Kaiwen, Cui Qi, etc Research on the Digital Generation of Miao Embroidery Patterns in Zunyi, Guizhou [J]. Journal of Donghua University: Social Sciences Edition, 2023, 23 (2): 49-58
  7. [7] CHEN H X，LI H X，LI Y H，et al. Multi-level Metric Learning for Few-Shot Image Recognition [C]//Artificial Neural Networks and Machine Learning(ICANN 2022).Bristol：Springer International Publishing，2022：243-254.
  8. [8] LI X，SONG K，LIN T，et al. PDAMeta: Meta-Learning Framework with Progressive Data Augmentation for Few-Shot Text Classification [C]//2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation(LREC-COLING 2024).Torino：ELRA and ICCL，2024：12668-12678.
  9. [9] SONG Y S，WANG T，MONDAL S K，et al. A Comprehensive Survey of Few-Shot Learning: Evolution, Applications, Challenges, and Opportunities [J/OL].arXiv:2205.06743 [cs.LG].（2022-05-13）.https://arxiv.org/abs/2205.06743v1.
  10. [10] TAN M X，LE Q V. EfficientNet: Rethinking Model Scaling for Convolutional Neural Networks [C]//International Conference on Machine Learning.Taiyuan：IMLS，2019：6105-6114.
  11. [11] SANDLER M，HOWARD A，ZHU M L，et al. MobileNetV2: Inverted Residuals and Linear Bottlenecks [C]//2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition.Salt Lake City：IEEE，2018：4510-4520.
  12. [12] HUANG G，LIU Z，MAATEN L V D，et al. Densely Connected Convolutional Networks [C]//2017 IEEE Conference on Computer Vision and Pattern Recognition（CVPR）.Honolulu：IEEE，2017：2261-2269.
  13. [13] SIMONYAN K，ZISSERMAN A. Very Deep Convolutional Networks for Large-scale Image Recognition [J/OL].arXiv:1409.1556 [cs.CV].（2014-09-04）.https://arxiv.org/abs/1409.1556.
  14. [14] KARRAS T，LAINE S，AITTALA M，et al. Analyzing and Improving the Image Quality of StyleGAN [C]//2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition(CVPR).Seattle：IEEE，2020：8107-8116.
  15. [15] KARRAS T，LAINE S，AILA T. A Style-based Generator Architecture for Generative Adversarial Networks [J].IEEE Transactions on Pattern Analysis and Machine Intelligence，2021，43（12）：4217-4228.
  16. [16] WANG Q L，WU B G，ZHU P F，et al. ECA-Net: Efficient Channel Attention for Deep Convolutional Neural Networks [C]//2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition(CVPR).Seattle：IEEE，2020：11531-11539.
  17. [17] Meng Wei, Yuan Yilin Review on the application of transfer learning in the diagnosis of novel coronavirus pneumonia [J]. Computer Science and Exploration, 2023,17 (3): 561-576
  18. [18] IMAN M，RASHEED K，ARABNIA H R. A Review of Deep Transfer Learning and Recent Advancements [J/OL].arXiv:2201.09679 [cs.LG].（2022-01-19）.https://arxiv.org/abs/2201.09679?context=cs.CV6.
  19. [19] TIAN S I P，REN Z，VENKATARAJ S，et al. Tackling Data Scarcity with Transfer Learning: A Case Study of thickness Characterization from Optical Spectra of Perovskite Thin Films [J/OL].arXiv:2207.02209 [cs.LG].（2022-06-14）.https://arxiv.org/abs/2207.02209.

Funding

Guizhou Province Science and Technology Plan Project (Qiankeheji-ZK [2021] General 340); Natural Science Research Project of Guizhou Provincial Department of Education (Qianjiaoji [2023] 061); Natural Science Research Project of Guizhou Provincial Department of Education (Qianjiaoji [2023] No. 012); Natural Science Research Project of Guizhou Provincial Department of Education (Qianjiaoji [2022] No. 047); Doctoral Research Launch Project of Guizhou University for Nationalities (GZMUZK [2024] QD11)

About the authors

Wu Jing (1988.05-), female, Miao ethnicity, from Guiyang, Guizhou, associate professor, master's supervisor, PhD, research direction: detection technology and automation devices; Yang Bangqin (1999.10-), female, Han ethnicity, from Liping, Guizhou. She is currently pursuing a master's degree in the field of massive data statistics and analysis; Zhang Yinjian (September 1998-), male, Han ethnicity, Guizhou Zhijin ethnic group, master's degree candidate, research direction: massive data statistics and analysis; Li Mingzhu (March 1983-), female, Han ethnicity, Hainan Linggao, associate professor, master's degree, research direction: intelligent detection and control, soft sensing modeling; Corresponding author: Chen Yan (September 1987-), female, Han ethnicity, from Zunyi, Guizhou, PhD candidate, research direction: Miao embroidery image recognition research and visualization.