Research on Automatic Recognition and Intelligent Recommendation System for Environmental Art Design Styles Based on Deep Learning
Main Article Content
Abstract
Automatic recognition and recommendation of environmental art design styles require robust visual feature extraction and personalized matching under blurred style boundaries and heterogeneous user preferences. To improve recognition accuracy and recommendation effectiveness, this study constructs a deep-learning-based style recognition and hybrid recommendation system. A dataset of 150,000 environmental art design images is built across 10 style categories, including modern, neoclassical, industrial, Nordic, Zen, neo-Baroque, pastoral, Mediterranean, minimalist, and postmodern styles. Images are cleaned, normalized, augmented, and annotated through expert voting. An improved ResNet-50 network with channel–spatial hybrid attention is designed to enhance style-related discriminative regions and capture both global stylistic semantics and local structural details. Stage-wise transfer learning is used for style classification, and deep 256-dimensional style embeddings are extracted for recommendation. User behavior logs, including browsing, liking, collecting, and forwarding, are weighted to construct long-term and short-term preference vectors. A hybrid recommendation strategy integrates style-similarity recall, item-based collaborative filtering secondary ranking, and diversity penalty adjustment. Experiments show that the improved model achieves 92.7% Top-1 accuracy and 98.4% Top-5 accuracy, outperforming VGG-16, Inception-v3, and original ResNet-50. The dynamic preference-fusion strategy improves click-through rate to 28.1% and diversity entropy to 1.82. The system provides an engineering framework for computer vision, image classification, and intelligent recommendation in visual design applications.
Downloads
Article Details
This work is licensed under a Creative Commons Attribution 4.0 International License.
Authors who publish with this journal agree to the following terms:
- Authors retain copyright and grant the journal right of first publication with the work simultaneously licensed under a Creative Commons Attribution License that allows others to share the work with an acknowledgement of the work's authorship and initial publication in this journal.
- Authors are able to enter into separate, additional contractual arrangements for the non-exclusive distribution of the journal's published version of the work (e.g., post it to an institutional repository or publish it in a book), with an acknowledgement of its initial publication in this journal.
- Authors are permitted and encouraged to post their work online (e.g., in institutional repositories or on their website) prior to and during the submission process, as it can lead to productive exchanges, as well as earlier and greater citation of published work (See The Effect of Open Access).
References
T. Imanishi, “A style classification based on evaluation experiments of residential interiors by experts and identification of explanatory elements and index: A preliminary study towards a classification system of interior styles,” Journal of the Human-Environment System, vol. 27, no. 1, pp. 9-22, 2025, doi: 10.1618/jhes.27.9.
H. Park B and H. Hyun K, “Analysis of pairings of colors and materials of furnishings in interior design with a data-driven framework,” Journal of Computational Design and Engineering, vol. 9, no. 6, pp. 2419-2438, 2022, doi: 10.1093/jcde/qwac114.
B. Wang, S. Zhang, J. Zhang, et al., “Architectural style classification based on CNN and channel-spatial attention,” Signal, Image and Video Processing, vol. 17, no. 1, pp. 99-107, 2023, doi: 10.1007/s11760-022-02208-0.
A. Torkashvand, M. S. Jameii, and A. Reza, “Deep learning-based collaborative filtering recommender systems: a comprehensive and systematic review,” Neural Computing and Applications, vol. 35, no. 33, pp. 24783-24827, 2023, doi: 10.1007/s00521-023-08958-3.
Z. Zhang, B. G. Patra, A. Yaseen, et al., “Scholarly recommendation systems: a literature survey,” Knowledge and Information Systems, vol. 65, no. 10, pp. 4433-4478, 2023, doi: 10.1007/s10115-023-01901-x.
G. Jain, T. Mahara, and C. Sharma S, “Effective time context based collaborative filtering recommender system inspired by Gower’s coefficient,” International Journal of System Assurance Engineering and Management, vol. 14, pp. 429-447, 2023, doi: 10.1007/s13198-022-01813-z.
J. Latrech, Z. Kodia, and N. Ben Azzouna, “CoDFi-DL: a hybrid recommender system combining enhanced collaborative and demographic filtering based on deep learning,” The Journal of Supercomputing, vol. 80, no. 1, pp. 1160-1182, 2024, doi: 10.1007/s11227-023-05519-2.
J. Jiang, H. Xu, X. Xu, et al., “Transformer-Based Fused Attention Combined with CNNs for Image Classification,” Neural Processing Letters, vol. 55, no. 9, pp. 11905-11919, 2023, doi: 10.1007/s11063-023-11402-1.
H. Bichri, A. Chergui, and M. Hain, “Image Classification with Transfer Learning Using a Custom Dataset: Comparative Study,” Procedia Computer Science, vol. 220, pp. 48-54, 2023, doi: 10.1016/j.procs.2023.03.009.
J. Wan, B. Li, K. Wang, et al., “An Improved ResNet50 for Environment Image Classification,” Procedia Computer Science, vol. 242, pp. 1000-1007, 2024, doi: 10.1016/j.procs.2024.08.246.
J. Zhang, X. Yu, X. Lei, et al., “A novel ResNet50-based attention mechanism for image classification,” Journal of Applied Science and Engineering, vol. 27, no. 8, pp. 2961-2969, 2024, doi: 10.6180/jase.202408_27(8).0004.
P. Su, X. Shen, H. Chen, et al., “M-AResNet: a novel multi-scale attention residual network for melting curve image classification,” Multimedia Tools and Applications, vol. 82, no. 27, pp. 42961-42976, 2023, doi: 10.1007/s11042-023-14694-6.
K. Singh, S. Dhawan, and N. Bali, “An Ensemble Learning Hybrid Recommendation System Using Content-Based, Collaborative Filtering, Supervised Learning and Boosting Algorithms,” Automatic Control and Computer Sciences, vol. 58, no. 5, pp. 491-505, 2024, doi: 10.3103/S0146411624700615.