Attention-guided transformer architecture for fine-grained style representation in design systems
Bao T. Nguyen, Thinh T. Nguyen · Journal of Computational Design and Engineering · 2026
Abstract Interior design style recognition presents a challenging computational problem due to fine-grained inter-style similarities, complex spatial arrangements, and strong variability in materials, lighting, and layout. From a computational design and engineering perspective, accurately modelling such stylistic variations is essential not only for classification, but also for enabling semantic representation, retrieval, and analysis of design cases within data-driven design systems. However, existing approaches often struggle to balance the extraction of fine-grained local cues with the modelling of global contextual relationships in complex interior scenes. In this study, we propose a unified data-model framework for interior design style recognition. First, we introduce InDes, a curated interior design image data set comprising 3826 images across five architectural styles. The data set integrates expert-labelled real-world samples from Vietnamese studios, with additional pseudo-labelled data generated through a semisupervised learning strategy, addressing class imbalance and supporting robust model training under limited annotation conditions. Second, we propose AT3 (Attention Transformer for Interior Style Design Classification), an attention-guided transformer architecture that combines convolutional feature extraction with hierarchical attention mechanisms and transformer-based global context modelling. By applying attention-based feature refinement prior to transformer tokenization, the proposed architecture preserves fine-grained stylistic cues while enabling effective long-range dependency modelling across interior scenes. This design improves the quality of learned representations for complex design environments characterized by subtle visual differences. Extensive experiments on the InDes data set demonstrate that AT3, when instantiated with an Xception backbone, achieves a validation accuracy of 87.8% and an F1-score of 0.86, outperforming strong convolutional and transformer-based baselines by margins of 5.2% and 6.7%, respectively. Ablation studies further validate the effectiveness of the proposed attention-guided architecture. Beyond classification performance, the learned representations provide a computational foundation for downstream design-oriented tasks such as style-aware retrieval, clustering, and data-driven analysis in computational design and engineering systems.