Deep Landscape Design Evaluation System with Multi-scale Visual Attention Mechanism

This research aims to explore and implement the application of multi-scale visual attention mechanism in a deep landscape design evaluation system. By integrating a deep learning model and a multi-scale visual attention mechanism, this study develops a new landscape design evaluation method that can automatically process and analyze a large number of landscape design images. The study first reviews the history and current development of landscape design, explores the theoretical basis of visual attention mechanisms, and analyzes in detail the potential application of multi-scale visual attention in landscape design. Furthermore, the application of deep learning in landscape design evaluation and its challenges are discussed in depth. In a case study, the developed model was used to evaluate selected urban parks, and the results showed that the model can effectively evaluate and improve the aesthetic appeal, functionality, environmental sustainability, and user satisfaction of landscape design. Finally, the discussion section highlights the significance and impact of the research findings, the advantages and limitations of the method, and the future application prospects of multi-scale visual attention mechanisms in the field of landscape design. This study provides a novel perspective on landscape design evaluation and demonstrates the potential of deep learning technology in urban planning and landscape design.