An AI-Assisted Evaluation Method for Assessing Hotel Brand Spatial Identity Alignment - Focusing on the Application of Brand Codebooks and Vision-Language Models -

Su-Min Hong; Jin-Kook Lee

doi:10.13161/kibim.2026.16.3.053

All Issue

2026 Vol.16, Issue 3 Preview Page

An AI-Assisted Evaluation Method for Assessing Hotel Brand Spatial Identity Alignment - Focusing on the Application of Brand Codebooks and Vision-Language Models - 호텔 브랜드 공간 아이덴티티 정합성 진단을 위한 AI 기반 보조 평가 방법 - 브랜드 코드북과 Vision-Language Model 활용을 중심으로 -

30 September 2026. pp. 53-67

PDF XML

Abstract

Recent advances in digital technologies and artificial intelligence (AI) have expanded the use of visual data in architectural and interior spatial analysis. However, the evaluation of qualitative design attributes, such as spatial atmosphere, material expression, and brand identity, remains largely dependent on expert judgment. Existing AI-based approaches have mainly focused on quantifying visual characteristics or assessing visual and stylistic similarity, while the structured evaluation of spatial images against multiple predefined qualitative design criteria remains less explored. This study proposes an AI-assisted method for evaluating hotel brand spatial identity alignment through three stages: (1) structuring hotel brand standards into brand-specific codebooks as fixed evaluation references, (2) evaluating spatial images using a Vision-Language Model (VLM) based on seven predefined criteria, and (3) validating the AI-based evaluation through a Human-in-the-Loop (HITL) comparison with professional evaluations. The dataset comprised 497 spatial images from 50 properties representing 12 hotel brands across four global hotel groups. For validation, 30 professionals independently evaluated 24 selected images using the same criteria and rating scale applied in the AI evaluation. The results showed a statistically significant positive correlation between AI and expert evaluations for overall identity coherence (Pearson’s r=.436, p=.033; Spearman’s ρ=.464, p=.014), with significant positive correlations observed in six of the seven evaluation criteria (r=.419–.459). However, the weighted kappa indicated limited absolute agreement (κ=.187), and localization showed no significant correlation (r=-.161, p=.453). These findings indicate that VLM-based evaluation can distinguish relative differences in brand spatial identity alignment when qualitative design criteria are structured and consistently applied, while also revealing limitations in absolute agreement and context-sensitive criteria. The proposed method provides a structured and repeatable framework for applying AI to qualitative spatial design evaluation and suggests its potential as a decision-support tool that complements, rather than replaces, expert judgment.

Keywords

Hotel Brand Spatial Identity

Brand Codebook

Vision-Language Model

AI-Assisted Evaluation

Human-in-the-Loop(HITL)

키워드

호텔 브랜드 공간 아이덴티티

브랜드 코드북

Vision-Language Model(VLM)

AI 기반 보조 평가

Human-in-the-Loop(HITL)

References

Anthropic (2026). Claude Sonnet 4.6 (claude-sonnet-4-6), https://docs.claude.com/en/docs/about-claude/models/overview (Sep. 27. 2026).

Chiang, C. H., Lee, H. Y. (2023). Can Large Language Models Be an Alternative to Human Evaluations?, Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pp. 15607-15631.

10.18653/v1/2023.acl-long.870

Choi, D. S., Jung, J. S. (2024). Visualization and Characteristic Analysis of Biophilic Design in Lifestyle Hotel Lobby Spaces Using Generative AI, Journal of Korea Society of Space Design, 19(4), pp. 60-69.

Cohen, J. (1968). Weighted Kappa: Nominal Scale Agreement with Provision for Scaled Disagreement or Partial Credit, Psychological Bulletin, 70(4), pp. 213-220.

10.1037/h0026256

de Roos, J. A. (2010). Hotel Management Contracts – Past and Present, Cornell Hospitality Quarterly, 51(1), pp. 68-80.

10.1177/1938965509354865

Han, Y., Lee, H. S. (2023). Analyzing Spatial Programs and Design Elements Using Deep Learning, Journal of the Architectural Institute of Korea, 39(9), pp. 105-114.

10.5659/JAIK.2023.39.9.105

Jeong, H., Lee, J. K. (2023). An Approach to Spatial Design Visualization through Design Style Combinations Based on Image-Generation AI, Journal of the Korean Housing Association, 34(6), pp. 17-24.

10.6107/JKHA.2023.34.6.017

Kim, Y. C., Lee, J. K. (2024). 360 Panorama Image-based Interior Architectural Visualization Data Archiving and its Applications, Journal of KIBIM, 14(4), pp. 1-9.

10.13161/KIBIM.2024.14.4.001

Landis, J. R., Koch, G. G. (1977). The Measurement of Observer Agreement for Categorical Data, Biometrics, 33(1), pp. 159-174.

10.2307/2529310

Lee, A. R., Heo, T. (2025). Exploring the Potential of LLM-Based AI to Replace Human Evaluation Tasks: Focusing on Building a Quality Monitoring System Using Expert AI, Information Systems Review, 27(4), pp. 203-227.

10.14329/isr.2025.27.4.203

Park, L., Yoo, J., Choi, Y. R., Oh, H. J. (2025). AI-Based Multimodal Analysis to Understand Design Languages: A Case Study of the Gaudi Architecture, Journal of Digital Contents Society, 26(5), pp. 1161-1169.

10.9728/dcs.2025.26.5.1161

Samjong KPMG Economic Research Institute (2025). Turning Point in the Hotel Industry: Business Trends and Growth Strategies, Business Focus, Samjong KPMG.

STR (2024). STR Chain Scales, https://www.costar.com/products/str-benchmark/resources/guidelines/str-chain-scales (Sep. 27. 2026).

Turner, M. J., Guilding, C. (2010). Hotel Management Contracts and Deficiencies in Owner-Operator Capital Expenditure Goal Congruency, Journal of Hospitality & Tourism Research, 34(4), pp. 478-511.

10.1177/1096348010370855

Yoo, Y., Kim, H., Lee, J. K. (2025). Automated BIM Modeling Based on AI-Generated Imagery - Focusing on Building Façade Wall Pattern Generation and Modeling -, Journal of KIBIM, 15(4), pp. 1-9.

Yoo, Y., Lee, J. K. (2024). Generative AI-based Exterior Building Design Visualization Approach in the Early Design Stage - Leveraging Architects' Style-trained Models -, Journal of KIBIM, 14(2), pp. 13-24.

10.13161/KIBIM.2024.14.2.013

Yoo, Y., Lee, J. K. (2025a). Quantification and Analysis Methods of Architectural Styles Using Image-Text AI Models, Journal of the Architectural Institute of Korea, 41(5), pp. 93-101.

Yoo, Y., Lee, J. K. (2025b). Development of Design Evaluation Approach and Tools Using Architectural Style Quantification System, Korean Journal of Computational Design and Engineering, 30(2), pp. 248-257.

10.7315/CDE.2025.248

Yu, E. S., Kim, G., Ahn, Y. H., Choi, J. S. (2024). A Methodology for Using ChatGPT to Improve BIM-based Design Data Evaluation System, Journal of KIBIM, 14(2), pp. 25-34.

10.13161/KIBIM.2024.14.2.025

Zheng, L., Chiang, W. L., Sheng, Y., Zhuang, S., Wu, Z., Zhuang, Y., Lin, Z., Li, Z., Li, D., Xing, E. P., Zhang, H., Gonzalez, J. E., Stoica, I. (2023). Judging LLM-as-a-Judge with MT-Bench and Chatbot Arena, Advances in Neural Information Processing Systems, 36, pp. 46595-46623.

10.52202/075280-2020

Information

Publisher :Journal of KIBIM
Publisher(Ko) :한국BIM학회
Journal Title :Journal of KIBIM
Journal Title(Ko) :한국BIM학회논문집
Volume : 16
No :3
Pages :53-67
Received Date : 2026-09-28; 2026-09-30
Accepted Date : 2026-09-30
DOI :https://doi.org/10.13161/kibim.2026.16.3.053

Journal of KIBIM ISSN:2288-1697(Print) 2288-1697(Online) 한국BIM학회논문집

All Issue