Lightweight Artistic Style Classification via Texture Correlation and Adaptive Local–Global Fusion
DOI:
https://doi.org/10.54097/hkbv4w72Keywords:
Artistic style classification, lightweight deep learning, texture correlation, adaptive feature fusion, contrastive learningAbstract
Artistic style classification remains challenging because visually similar styles often exhibit subtle differences in brushstrokes, textures, colors, and global composition. To address this problem, this paper proposes a lightweight artistic style classification framework based on a MobileNet–Transformer Hybrid (MTH). First, a Lightweight Texture Correlation Module (LTCM) is introduced to capture multi-scale texture relationships from intermediate convolutional features. Second, an Adaptive Local–Global–Texture Fusion module (ALGTF) dynamically integrates local details, global composition, and texture representations according to their relative contributions. Third, a Confusable-Style Hard Contrastive Learning strategy (CSHCL) is designed to emphasize difficult inter-style samples and improve feature separability between visually similar artistic categories. The Dynamic Channel-Spatial Attention Module (DCSAM) is retained to enhance style-sensitive regions and channels. Experiments on ArtBench-10 and Painting91_Styles demonstrate that the proposed framework provides improved classification performance while maintaining a lightweight architecture, showing its effectiveness for fine-grained artistic image analysis.
Downloads
References
[1] Ma, K., Lee, S., Ma, X., & Chen, H. (2025). Fine art image classification and design methods integrating lightweight deep learning. Scientific Reports, 15, Article 33006. https://doi.org/10.1038/s41598-025-18420-0
[2] Wang, Q., & Feng, G. (2025). Multimodal style aggregation network for art image classification. Signal Processing: Image Communication, 137, Article 117309. https://doi.org/10.1016/j.image.2025.117309
[3] Liang, S., Pan, L., Mohammadzadeh, F., et al. (2025). Enhancing artistic style classification through a novel ArtFusionNet framework. Scientific Reports, 15, Article 36178. https://doi.org/10.1038/s41598-025-20087-6
[4] Wang, M., Xia, F., Yang, T., Zou, Y., & Zhou, Y. (2026). An art style classification network integrating contrastive learning and counterfactual attention. PLOS ONE, 21(6), Article e0343954. https://doi.org/10.1371/journal.pone.0343954
[5] Xiao, B., & Li, H. (2025). Painting style classification and art image analysis model by fusion of convolutional neural network and principal component analysis. Journal of Computational Methods in Sciences and Engineering, 25(5), 4048–4060. https://doi.org/10.1177/14727978251337910
[6] Xu, B., & Lu, X. (2026). Artistic style classification in traditional cultural heritage using convolutional neural networks. Scientific Reports, 16, Article 20389. https://doi.org/10.1038/s41598-026-49996-w
[7] Zhang, H. (2026). Deep learning approach for art style recognition and colour restoration in visual art images. Scientific Reports, 16, Article 24487. https://doi.org/10.1038/s41598-026-51895-z
[8] Li, W. (2025). Enhanced automated art curation using supervised modified CNN for art style classification. Scientific Reports, 15, Article 7319. https://doi.org/10.1038/s41598-025-91671-z
[9] El Vaigh, C. B., Garcia, N., Renoust, B., et al. (2025). GNNBoost: Boosting artwork classification with graph embeddings. Multimedia Tools and Applications, 84, 39353–39373. https://doi.org/10.1007/s11042-025-20699-0
[10] Elkin, G., Shahar, O. I., Ohayon, Y., Alali, N., & Ben-Shahar, O. (2025). Recognizing artistic style of archaeological image fragments using deep style extrapolation. In Culture and Computing, HCII 2025 (Lecture Notes in Computer Science, Vol. 15800, pp. 115–131). https://doi.org/10.1007/978-3-031-93160-4_8
[11] Luo, M., Liu, L., Lu, Y., & Suen, C. Y. (2025). Art style classification via self-supervised dual-teacher knowledge distillation. Applied Soft Computing, 174, Article 112964. https://doi.org/10.1016/j.asoc.2025.112964
[12] Liu, Y., Bai, H., & Wang, J. (2024). Fine-art recognition using convolutional transformers. PeerJ Computer Science, 10, Article e2409. https://doi.org/10.7717/peerj-cs.2409
[13] Lee, S., Choi, J., & Kim, H. J. (2025). EfficientViM: Efficient vision Mamba with hidden state mixer based state space duality. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (pp. 14923–14933).
[14] He, H., Zhang, J., Cai, Y., et al. (2025). MobileMamba: Lightweight multi-receptive visual Mamba network. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition.
[15] Zhang, J., Hu, T., He, H., et al. (2025). EMOv2: Pushing 5M vision model frontier. IEEE Transactions on Pattern Analysis and Machine Intelligence, 47(11), 10560–10576. https://doi.org/10.1109/TPAMI.2025.3596776
[16] Tschannen, M., Gritsenko, A., Wang, X., et al. (2025). SigLIP 2: Multilingual vision-language encoders with improved semantic understanding, localization, and dense features. arXiv preprint arXiv:2502.14786. https://doi.org/10.48550/arXiv.2502.14786
[17] Siméoni, O., Vo, H. V., Seitzer, M., et al. (2025). DINOv3. arXiv preprint arXiv:2508.10104. https://doi.org/10.48550/arXiv.2508.10104
[18] Balmaseda, V., Wang, B., Lin, C.-L., & Yang, T. (2025). Discovering global false negatives on the fly for self-supervised contrastive learning. In Proceedings of the 42nd International Conference on Machine Learning (Proceedings of Machine Learning Research, Vol. 267, pp. 2697–2714).
[19] Zhou, Z.-H., Wang, J.-J., Wei, T., & Zhang, M.-L. (2025). Weakly-supervised contrastive learning for imprecise class labels. In Proceedings of the 42nd International Conference on Machine Learning (Proceedings of Machine Learning Research, Vol. 267, pp. 79328–79365).
[20] Hieu, N. M., & Ledent, A. (2025). Generalization analysis for supervised contrastive representation learning under non-IID settings. In Proceedings of the 42nd International Conference on Machine Learning (Proceedings of Machine Learning Research, Vol. 267, pp. 23179–23218).
Downloads
Published
Issue
Section
License
Copyright (c) 2026 International Journal of Advanced Engineering and Technology Research
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.
