Detecting Traditional Chinese Window Patterns in Suzhou Gardens Using YOLOv8: A Computer Vision Approach to Architectural Heritage Recognition

Jiangyang Zhao; Jingyang Liu; Davide Lombardi

Detecting Traditional Chinese Window Patterns in Suzhou Gardens Using YOLOv8: A Computer Vision Approach to Architectural Heritage Recognition

Department of Architecture

Research output: Chapter in Book or Report/Conference proceeding › Conference Proceeding › peer-review

Abstract

Traditional lattice windows in Suzhou gardens, core elements of Chinese architectural heritage, face identification challenges due to intricate patterns and inefficient manual methods. This study develops a YOLOv8n-based approach for automated detection of thirteen lattice types (documented in Yuanye and Yinzhao Fayuan), using a 488-image field dataset. The model achieves mAP50 = 0.7299, effective under controlled conditions. Integrated into H-Scan software with DeepSeek LLM, it provides cultural context (symbolism, craftsmanship) and bilingual interaction, aiding students, tourists, and researchers. Validating lightweight AI for heritage recognition, this work bridges technical detection and digital education, supporting architectural knowledge preservation.

Original language	English
Title of host publication	Humanistic Computation and Intelligence - Proceedings of the 31st International Conference on Computer-Aided Architectural Design Research in Asia, CAADRIA 2026 - Volume 3
Editors	Mohammed Makki, Jielin Chen, Sheng-Fen Nik Chien, Jie Liu, Likai Wang, Shih-Yuan Wang
Publisher	The Association for Computer-Aided Architectural Design Research in Asia
Pages	21-30
Number of pages	10
ISBN (Print)	9789887135906
Publication status	Published - 2026
Event	31st International Conference on Computer-Aided Architectural Design Research in Asia, CAADRIA 2026 - Hsinchu, Taiwan, Province of China Duration: 26 Apr 2026 → 2 May 2026

Publication series

Name	Proceedings of the International Conference on Computer-Aided Architectural Design Research in Asia
Volume	3
ISSN (Print)	2710-4257
ISSN (Electronic)	2710-4265

Conference

Conference	31st International Conference on Computer-Aided Architectural Design Research in Asia, CAADRIA 2026
Country/Territory	Taiwan, Province of China
City	Hsinchu
Period	26/04/26 → 2/05/26

Keywords

Architectural Heritage
Computer Vision
Object Detection
Suzhou Garden
Traditional Chinese Window

Cite this

Zhao, J., Liu, J., & Lombardi, D. (2026). Detecting Traditional Chinese Window Patterns in Suzhou Gardens Using YOLOv8: A Computer Vision Approach to Architectural Heritage Recognition. In M. Makki, J. Chen, S.-F. Nik Chien, J. Liu, L. Wang, & S.-Y. Wang (Eds.), Humanistic Computation and Intelligence - Proceedings of the 31st International Conference on Computer-Aided Architectural Design Research in Asia, CAADRIA 2026 - Volume 3 (pp. 21-30). (Proceedings of the International Conference on Computer-Aided Architectural Design Research in Asia; Vol. 3). The Association for Computer-Aided Architectural Design Research in Asia.

Zhao, Jiangyang ; Liu, Jingyang ; Lombardi, Davide. / Detecting Traditional Chinese Window Patterns in Suzhou Gardens Using YOLOv8 : A Computer Vision Approach to Architectural Heritage Recognition. Humanistic Computation and Intelligence - Proceedings of the 31st International Conference on Computer-Aided Architectural Design Research in Asia, CAADRIA 2026 - Volume 3. editor / Mohammed Makki ; Jielin Chen ; Sheng-Fen Nik Chien ; Jie Liu ; Likai Wang ; Shih-Yuan Wang. The Association for Computer-Aided Architectural Design Research in Asia, 2026. pp. 21-30 (Proceedings of the International Conference on Computer-Aided Architectural Design Research in Asia).

@inproceedings{5c8f9e9c8480417b89876df9b59c4b57,

title = "Detecting Traditional Chinese Window Patterns in Suzhou Gardens Using YOLOv8: A Computer Vision Approach to Architectural Heritage Recognition",

abstract = "Traditional lattice windows in Suzhou gardens, core elements of Chinese architectural heritage, face identification challenges due to intricate patterns and inefficient manual methods. This study develops a YOLOv8n-based approach for automated detection of thirteen lattice types (documented in Yuanye and Yinzhao Fayuan), using a 488-image field dataset. The model achieves mAP50 = 0.7299, effective under controlled conditions. Integrated into H-Scan software with DeepSeek LLM, it provides cultural context (symbolism, craftsmanship) and bilingual interaction, aiding students, tourists, and researchers. Validating lightweight AI for heritage recognition, this work bridges technical detection and digital education, supporting architectural knowledge preservation.",

keywords = "Architectural Heritage, Computer Vision, Object Detection, Suzhou Garden, Traditional Chinese Window",

author = "Jiangyang Zhao and Jingyang Liu and Davide Lombardi",

note = "Publisher Copyright: {\textcopyright} 2026 and published by the Association for Computer-Aided Architectural Design Research in Asia (CAADRIA), Hong Kong.; 31st International Conference on Computer-Aided Architectural Design Research in Asia, CAADRIA 2026 ; Conference date: 26-04-2026 Through 02-05-2026",

year = "2026",

language = "English",

isbn = "9789887135906",

series = "Proceedings of the International Conference on Computer-Aided Architectural Design Research in Asia",

publisher = "The Association for Computer-Aided Architectural Design Research in Asia",

pages = "21--30",

editor = "Mohammed Makki and Jielin Chen and \{Nik Chien\}, Sheng-Fen and Jie Liu and Likai Wang and Shih-Yuan Wang",

booktitle = "Humanistic Computation and Intelligence - Proceedings of the 31st International Conference on Computer-Aided Architectural Design Research in Asia, CAADRIA 2026 - Volume 3",

}

Zhao, J , Liu, J & Lombardi, D 2026, Detecting Traditional Chinese Window Patterns in Suzhou Gardens Using YOLOv8: A Computer Vision Approach to Architectural Heritage Recognition. in M Makki, J Chen, S-F Nik Chien, J Liu, L Wang & S-Y Wang (eds), Humanistic Computation and Intelligence - Proceedings of the 31st International Conference on Computer-Aided Architectural Design Research in Asia, CAADRIA 2026 - Volume 3. Proceedings of the International Conference on Computer-Aided Architectural Design Research in Asia, vol. 3, The Association for Computer-Aided Architectural Design Research in Asia, pp. 21-30, 31st International Conference on Computer-Aided Architectural Design Research in Asia, CAADRIA 2026, Hsinchu, Taiwan, Province of China, 26/04/26.

Detecting Traditional Chinese Window Patterns in Suzhou Gardens Using YOLOv8: A Computer Vision Approach to Architectural Heritage Recognition. / Zhao, Jiangyang ; Liu, Jingyang ; Lombardi, Davide.
Humanistic Computation and Intelligence - Proceedings of the 31st International Conference on Computer-Aided Architectural Design Research in Asia, CAADRIA 2026 - Volume 3. ed. / Mohammed Makki; Jielin Chen; Sheng-Fen Nik Chien; Jie Liu; Likai Wang; Shih-Yuan Wang. The Association for Computer-Aided Architectural Design Research in Asia, 2026. p. 21-30 (Proceedings of the International Conference on Computer-Aided Architectural Design Research in Asia; Vol. 3).

Research output: Chapter in Book or Report/Conference proceeding › Conference Proceeding › peer-review

TY - GEN

T1 - Detecting Traditional Chinese Window Patterns in Suzhou Gardens Using YOLOv8

T2 - 31st International Conference on Computer-Aided Architectural Design Research in Asia, CAADRIA 2026

AU - Zhao, Jiangyang

AU - Liu, Jingyang

AU - Lombardi, Davide

PY - 2026

Y1 - 2026

N2 - Traditional lattice windows in Suzhou gardens, core elements of Chinese architectural heritage, face identification challenges due to intricate patterns and inefficient manual methods. This study develops a YOLOv8n-based approach for automated detection of thirteen lattice types (documented in Yuanye and Yinzhao Fayuan), using a 488-image field dataset. The model achieves mAP50 = 0.7299, effective under controlled conditions. Integrated into H-Scan software with DeepSeek LLM, it provides cultural context (symbolism, craftsmanship) and bilingual interaction, aiding students, tourists, and researchers. Validating lightweight AI for heritage recognition, this work bridges technical detection and digital education, supporting architectural knowledge preservation.

AB - Traditional lattice windows in Suzhou gardens, core elements of Chinese architectural heritage, face identification challenges due to intricate patterns and inefficient manual methods. This study develops a YOLOv8n-based approach for automated detection of thirteen lattice types (documented in Yuanye and Yinzhao Fayuan), using a 488-image field dataset. The model achieves mAP50 = 0.7299, effective under controlled conditions. Integrated into H-Scan software with DeepSeek LLM, it provides cultural context (symbolism, craftsmanship) and bilingual interaction, aiding students, tourists, and researchers. Validating lightweight AI for heritage recognition, this work bridges technical detection and digital education, supporting architectural knowledge preservation.

KW - Architectural Heritage

KW - Computer Vision

KW - Object Detection

KW - Suzhou Garden

KW - Traditional Chinese Window

UR - https://www.scopus.com/pages/publications/105048874588

M3 - Conference Proceeding

AN - SCOPUS:105048874588

SN - 9789887135906

T3 - Proceedings of the International Conference on Computer-Aided Architectural Design Research in Asia

SP - 21

EP - 30

BT - Humanistic Computation and Intelligence - Proceedings of the 31st International Conference on Computer-Aided Architectural Design Research in Asia, CAADRIA 2026 - Volume 3

A2 - Makki, Mohammed

A2 - Chen, Jielin

A2 - Nik Chien, Sheng-Fen

A2 - Liu, Jie

A2 - Wang, Likai

A2 - Wang, Shih-Yuan

PB - The Association for Computer-Aided Architectural Design Research in Asia

Y2 - 26 April 2026 through 2 May 2026

ER -

Zhao J , Liu J , Lombardi D. Detecting Traditional Chinese Window Patterns in Suzhou Gardens Using YOLOv8: A Computer Vision Approach to Architectural Heritage Recognition. In Makki M, Chen J, Nik Chien SF, Liu J, Wang L, Wang SY, editors, Humanistic Computation and Intelligence - Proceedings of the 31st International Conference on Computer-Aided Architectural Design Research in Asia, CAADRIA 2026 - Volume 3. The Association for Computer-Aided Architectural Design Research in Asia. 2026. p. 21-30. (Proceedings of the International Conference on Computer-Aided Architectural Design Research in Asia).

Detecting Traditional Chinese Window Patterns in Suzhou Gardens Using YOLOv8: A Computer Vision Approach to Architectural Heritage Recognition

Abstract

Publication series

Conference

Keywords

Other files and links

Reinterpreting Traditional Windows Lattices Patterns in Contemporary Design through Parametric Design and Digital Fabrication

SURF-2025-0165, Reinterpreting Traditional Windows Lattices Patterns in Contemporary Design through Parametric Design and Digital Fabrication

Cite this

Detecting Traditional Chinese Window Patterns in Suzhou Gardens Using YOLOv8: A Computer Vision Approach to Architectural Heritage Recognition

Abstract

Publication series

Conference

Keywords

Other files and links

Projects

Reinterpreting Traditional Windows Lattices Patterns in Contemporary Design through Parametric Design and Digital Fabrication

Activities

SURF-2025-0165, Reinterpreting Traditional Windows Lattices Patterns in Contemporary Design through Parametric Design and Digital Fabrication

Cite this