Skip to main content

ORIGINAL RESEARCH article

Front. Built Environ.
Sec. Building Information Modelling (BIM)
Volume 10 - 2024 | doi: 10.3389/fbuil.2024.1334704

Enhancing Image Fusion and Quality Improvement in Modern Landscape Architecture Style Design using CycleGAN with Attention Mechanism optimized by ResNet-50 Provisionally Accepted

 Chi Gao1* Junlei Zhang1
  • 1Huazhong Agricultural University, China

The final, formatted version of the article will be published soon.

Receive an email when it is updated
You just subscribed to receive the final version of the article

Traditional landscape design heavily relies on the experience and imagination of designers.Incorporating modern architectural styles can provide designers with more creativity and inspiration, making the design more innovative and unique. We need a method to introduce elements of modern architectural style and provide visual references during the design process.The challenge we face is how to integrate modern architectural styles into landscape design to enhance inspiration and creativity.This paper aims to leverage advanced computer vision techniques, particularly CycleGAN and attention mechanism optimized with ResNet-50, to enhance image fusion and improve image quality while incorporating elements of modern landscape architectural style. We first collected a dataset containing photographs of modern landscape architecture and various style reference images. Next, we utilized the CycleGAN model to transform the modern landscape architecture photos into reference images of the target style. Subsequently, we introduced an attention mechanism to enhance the perception of modern landscape architectural styles by optimizing the ResNet-50 model. To fine-tune the model, we employed the generated images as training data and inputted them into the optimized ResNet-50 model for transfer learning. Finally, we performed image fusion between the outputs of the generator and the input images. We applied this model to practical modern landscape architecture style design and generated reference images to guide the design process. By combining these techniques, our goal is to achieve image fusion and quality improvement.We conducted a series of experiments to evaluate the effectiveness of the proposed method. The experimental results demonstrate significant improvements in image fusion and quality enhancement. By incorporating elements of modern landscape architectural style into image processing, we 1 Zhanget al.enhance the perceptual effect of the images, presenting more appealing and aesthetic visual results to observers. The significance of thispaper article lies in providing a novel and effective approach to image fusion and quality improvement, offering valuable insights for research and applications in the field of computer vision and image processing. By incorporating elements of modern architectural styles into landscape design, designers can gain more creativity and inspiration, resulting in more innovative and unique designs.

Keywords: CycleGAN, attention mechanism, Resnet-50, image fusion, Quality Improvement, modern landscape architectural style

Received: 07 Nov 2023; Accepted: 23 Jan 2024.

Copyright: © 2024 Gao and Zhang. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

* Correspondence: Mx. Chi Gao, Huazhong Agricultural University, Wuhan, China