Deep Learning and Computer Vision-Based Art Style Evaluation Model for Digital Painting
International Journal of e-Collaboration · 12.11.2026 · 22 (1) · 1-37
DOI: 10.4018/ijec.427556 ↗Kurz gefasst
Fachartikel in der Zeitschrift International Journal of e-Collaboration, erschienen am 12.11.2026. Verfasst von Wei Tang und 3 weiteren Autor:innen. Die Arbeit nennt 29 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 3.0).
Abstract (Originalsprache)
This research proposes an Adaptive Differential Evolution-Vision Transformer-Convolutional Neural Network (AdapDE-VTrans-CNNet) for intelligent art style evaluation. Image preprocessing includes normalization and data augmentation to improve diversity and generalization. Histogram of Oriented Gradients (HOG) extracts structural and texture features, while the hybrid CNN-Vision Transformer captures complementary local and global representations. The CNNet employs hierarchical feature extraction with attention layers blocks to refine stylistically discriminative features. The VTrans captures long-range dependencies and global artistic characteristics, supported by Grad-CAM, Grad-CAM++, and Layer-wise Relevance Propagation (LRP) for model interpretability. Adaptive Differential Evolution optimizes network parameters for faster convergence. The proposed model achieved 96.4% accuracy, 95.1% precision, 95.8% recall, and 95.4% F1-score, demonstrating accurate and scalable digital artwork style evaluation.
Zitieren
APA
Tang, W., Yu, D., Zhang, T., & Li, J. (2026). Deep Learning and Computer Vision-Based Art Style Evaluation Model for Digital Painting. International Journal of e-Collaboration, 22(1), 1-37. https://doi.org/10.4018/ijec.427556
Vancouver
Tang W, Yu D, Zhang T, Li J. Deep Learning and Computer Vision-Based Art Style Evaluation Model for Digital Painting. International Journal of e-Collaboration. 2026;22(1):1-37. doi:10.4018/ijec.427556
BibTeX
@article{Tang2026,
author = {Tang, Wei and Yu, Defeng and Zhang, Tao and Li, Jiaxue},
title = {Deep Learning and Computer Vision-Based Art Style Evaluation Model for Digital Painting},
journal = {International Journal of e-Collaboration},
year = {2026},
volume = {22},
number = {1},
pages = {1--37},
doi = {10.4018/ijec.427556}
}
EUVIMED