Innovative Framework for Historical Architectural Recognition in China: Integrating Swin Transformer and Global Channel–Spatial Attention Mechanism

Buildings, Journal Year: 2025, Volume and Issue: 15(2), P. 176 - 176

Published: Jan. 9, 2025

The digital recognition and preservation of historical architectural heritage has become a critical challenge in cultural inheritance sustainable urban development. While deep learning methods show promise classification, existing models often struggle to achieve ideal results due the complexity uniqueness buildings, particularly limited data availability remote areas. Focusing on study Chinese architecture, this research proposes an innovative framework that integrates Swin Transformer backbone with custom-designed Global Channel Spatial Attention (GCSA) mechanism, thereby substantially enhancing model’s capability extract details comprehend global contextual information. Through extensive experiments constructed building dataset, our model achieves outstanding performance over 97.8% key metrics including accuracy, precision, recall, F1 score (harmonic mean precision recall), surpassing traditional CNN (convolutional neural network) architectures contemporary models. To gain deeper insights into decision-making process, we employed comprehensive interpretability t-SNE (t-distributed Stochastic Neighbor Embedding), Grad-CAM (gradient-weighted class activation mapping), multi-layer feature map analysis, revealing systematic extraction process from structural elements material textures. This offers substantial technical support for modeling establishing foundation damage assessment. It contributes formulation precise restoration strategies provides scientific basis governments institutions develop region-specific policies conservation efforts.

Language: Английский

Landscape Visual Affordance Evaluation at a Regional Scale in National Parks: A Case Study of the Changhong Area in Qianjiangyuan National Park

Land, Journal Year: 2025, Volume and Issue: 14(3), P. 589 - 589

Published: March 11, 2025

National parks play a vital role in safeguarding natural scenery, maintaining ecological integrity, and preserving cultural heritage, while simultaneously offering valuable opportunities for recreation education. Among the diverse resources provided by national parks, visual landscape hold particular significance due to their capacity inspire, educate, enhance aesthetic appreciation. However, assessing managing these remain challenging, as they span both physical attributes of human perception process. This study aims develop theoretical practical framework evaluating “landscape affordance” parks. Grounded psychology’s affordance theory, proposed approach integrates sensory affordance, encompassing objective subjective processes perception. Drawing on multi-dimensional set indicators, research quantifies features—such topography, land use, vegetation cover, structure—as well dimensions, including visibility, prominence, viewing frequency. These elements are synthesized into visibility assessment model built upon theory framework. The results demonstrate that effectively identifies patches with varying degrees quality importance within other protected areas. By providing robust support management decisions—such zoned protection, optimizing recreational facilities, visitor carrying capacity—this offers new insights guidance sustainable planning landscapes ecologically critical regions.

Language: Английский

Designing effective image-based surveys for urban visual perception

Landscape and Urban Planning, Journal Year: 2025, Volume and Issue: 260, P. 105368 - 105368

Published: April 17, 2025

Language: Английский