[JCR] CGAN-Assisted Museum Architecture

The Floor Plan Design Method of Exhibition Halls in CGAN-Assisted Museum Architecture

Abstract: The paper proposes a method for designing the floor plans of museum exhibition halls using a conditional generative adversarial network (CGAN). The traditional approach of designing exhibition hall floor plans individually for each floor in a multi-story building is time-consuming and inefficient. The CGAN-based method aims to streamline the design process and help architects work more efficiently.

The paper introduces the basic concepts and structures of CGAN. CGAN is a machine learning model with a generator network and a discriminator network. The generator network generates samples based on input conditions, while the discriminator network distinguishes real from generated samples. By training these networks together adversarially, the CGAN learns to generate realistic samples matching the desired conditions.

The design and training process of the CGAN model for museum exhibition hall floor plans are described in detail. The paper also briefly discusses the datasets used for training and the evaluation metrics employed.

In the Results and Discussion sections, the paper presents an example of a generated floor plan design for a museum exhibition hall. The design is evaluated and analyzed. The advantages of the method are summarized, including generating floor plans with regularity based on given conditions, diverse plan designs for different schemes, and design optimization through human-computer interaction. The iterative improvement of the design according to user needs enhances practicality.

However, the paper acknowledges limitations. One is the need for a sufficient dataset. Insufficient data limits the scope of application, e.g., museums converted from certain historical buildings.

In conclusion, the proposed CGAN-based method generates museum exhibition hall floor plan designs with regularity, providing flexibility and optimization through human-computer interaction. Adequate datasets are crucial, and limitations exist. [SCITIP CGAN-Assisted Museum]

DOI:  https://doi.org/10.3390/buildings13030756

CITE: Min, Xiao, Liang Zheng, and Yile Chen. 2023. “The Floor Plan Design Method of Exhibition Halls in CGAN-Assisted Museum Architecture” Buildings 13, no. 3: 756. https://doi.org/10.3390/buildings13030756

作者的话

1. 心得分享

今日分享之论文，是本月见刊上线的第一篇 SCIE 期刊论文，我认为很有很有必要纪念一下，因为这篇论文不仅意味着我从一个新的建筑类型进行探索（有别于之前做过的熟悉的类型），很多东西要想的时候，还是要重新弄懂。还有一个原因就是，这篇论文有 6 位审稿人！！！！！而且修改了 3 轮（大修一轮，小修一轮，后面编辑叫改了题目又是小修一轮），我从来没试过这样。但也可以自信地认为，这篇论文经过了 6 位审稿人的意见修改，质量还是不错的，写出来的东西，还是有意义的，有阅读价值的。

言归正传，这篇论文《The Floor Plan Design Method of Exhibition Halls in CGAN-assisted Museum Architecture》（CGAN 辅助博物馆建筑展厅平面图设计方法研究）也是去年暑假就开始着手，但收集材料是花了巨大的时间。本人并非第一作者，为本文的通讯作者，在本次论文中主要是撰写第一章至第二章前半部分、全文翻译（含改写）、写 coverletter、联系客座编辑、返修了三轮、投稿等。欢迎大家多多阅读呀！支持一下我们家潇潇人生中第一篇 SCI😀

线上阅读地址：https://www.mdpi.com/2075-5309/13/3/756

2. 简单讲讲

利用 CGAN 训练学习一堆博物馆展厅的平面图，学会了它的特征（展厅平面的柜子、布置、流线等）以后，尝试给它一个轮廓是否可以生成展厅平面的设计，最后建立了 3D 视图的模式，并且与人工设计的对比，总结优缺点。

文章大纲标题：

1. Introduction

1.1. Research Background: The Importance and Research Status of Floor Plan Design of Museum Exhibition Halls

1.2. Literature Review

1.3. Problem Statement and Objectives

2. Methods and Museum Situation Analysis

2.1. Research Methods and Process

2.2. Material Handling

2.3. CGAN Model

3. Model Training Process and Results

3.1. Training Process

3.2. Model Test

4. Discussion: Model Application and Comparison

4.1. Model Application

4.2. Comparative Analysis of Differences with Architects’Designs

4.3. Generating a Variety of Museum Exhibition Hall Design Schemes

5. Conclusions

Appendix A

References

3. 文章插图