COMPARISON OF AI-GENERATED CONCEPTUAL INTERIOR DESIGNS USING STANDARDIZED PROMPTS IN VARIOUS AI MODELS

Yıl 2025, Cilt: 15 Sayı: 4, 1365 - 1386, 01.10.2025

Mehmet Uğur Kahraman , Yaren Şekerci , Enes Can Kılıç

https://doi.org/10.7456/tojdac.1736051

Öz

This study evaluates the performance of three widely-used text-to-image AI models—DALL-E, MidJourney, and Stable Diffusion—in generating interior design visuals. Five standardized prompts, covering essential design elements such as architectural features, style, furniture, lighting, textiles, accessories, and layout, were provided to each AI model. The resulting visuals were then assessed by 15 experienced professionals and academics with over 10 years of expertise in interior design. Using a detailed rating system, the study found that MidJourney consistently outperformed the others, excelling in categories like design style, textiles, and lighting. DALL-E proved effective in generating architectural features but struggled with finer details, particularly in textiles and lighting. Stable Diffusion lagged behind in overall performance, especially in categories like layout and furniture. The study highlights the importance of prompt specificity in shaping AI-generated visuals and suggests areas for future improvement in AI models, particularly in complex design tasks.

Anahtar Kelimeler

AI-generated interior visuals , Text-to-image AI models , Interior design evaluation , AI in design visualization , AI models comparison

Etik Beyan

Before commencing the experiment, which is the chosen research method, the Antalya Bilim University Natural and Applied Sciences Scientific Research and Publication Ethics Committee, in their decision dated September 12, 2024, and marked as 2024/05, unanimously agreed that the applicant bears the responsibility for the research's methodology and scope, in alignment with ethical, legal, and copyright considerations, making it ethically sound to carry out the study.

FARKLI YAPAY ZEKÂ MODELLERİ KULLANILARAK STANDART KOMUTLARLA ÜRETİLEN KAVRAMSAL İÇ MEKÂN TASARIMLARININ KARŞILAŞTIRILMASI

Öz

Bu çalışma, metinden görsele dönüştürme yeteneğine sahip üç yaygın yapay zekâ modeli olan DALL-E, MidJourney ve Stable Diffusion’ın iç mekân tasarımı görselleri üretme performansını değerlendirmektedir. Her bir yapay zekâ modeline, mimari özellikler, stil, mobilya, aydınlatma, tekstil, aksesuarlar ve yerleşim gibi temel tasarım öğelerini kapsayan beş standart komut verilmiştir. Ortaya çıkan görseller, iç mekân tasarımı alanında 10 yılı aşkın deneyime sahip 15 profesyonel ve akademisyen tarafından değerlendirilmiştir. Detaylı bir puanlama sistemi kullanılarak yapılan değerlendirme sonucunda, MidJourney’nin özellikle tasarım stili, tekstil ve aydınlatma gibi kategorilerde üstün performans sergileyerek diğer modelleri geride bıraktığı belirlenmiştir. DALL-E, mimari öğeleri üretmede etkili olurken, özellikle tekstil ve aydınlatma gibi ince detaylarda yetersiz kalmıştır. Stable Diffusion ise genel performans açısından geride kalmış; özellikle yerleşim düzeni ve mobilya kategorilerinde düşük başarı göstermiştir. Çalışma, yapay zekâ tarafından üretilen görsellerin niteliğinde komutların özgüllüğünün önemine dikkat çekmekte ve karmaşık tasarım görevlerinde yapay zekâ modellerinin gelişime açık alanlarını ortaya koymaktadır.

Anahtar Kelimeler

Yapay zekâ üretimi iç mekân görselleri , Metinden görsele yapay zekâ modell , İç mekân tasarımı değerlendirmesi , Tasarım görselleştirmede yapay zekâ , Yapay zekâ modellerinin karşılaştırılması

Kaynakça

• Albaghajati, Z. M., Bettaieb, D. M., & Malek, R. B. (2023). Exploring text-to-image application in architectural design: Insights and implications. Architecture, Structures and Construction, 3, 475–497.
• Bertani, F., Ponta, L., Raberto, M., Teglio, A., & Cincotti, S. (2021). The complexity of the intangible digital economy: An agent-based model. Journal of Business Research, 527-540.
• Betker, J., Goh, G., Jing, L., Brooks, T., Wang, J., Li, L., & Ramesh, A. (2023). Improving image generation with better captions. Computer Science, 2(3), 8.
• Bongomin, O., Gilibrays Ocen, G., Oyondi Nganyi, E., Musinguzi, A., & Omara, T. (2020). Exponential disruptive technologies and the required skills of industry 4.0. Journal of Engineering, 2020, 4280156.
• Brisco, R., Hay, L., & Dhami, S. (2023). Exploring the role of text-to-image AI in concept generation. Proceedings of the Design Society, 3, 1835–1844.
• Caan, S. (2011). Rethinking design and interiors: Human beings in the built environment. Hachette UK.
• Caires, C. S., Estadieu, G., & Olga Ng Ka Man, S. (2023). Design thinking methodology and text-to-image artificial intelligence: A case study in the context of furniture design education. In Perspectives on Design and Digital Communication IV: Research, Innovations and Best Practices (pp. 113–134). Springer.
• Chen, X., Xie, H., & Hwang, G.-J. (2020). A multi-perspective study on artificial intelligence in education: Grants, conferences, journals, software tools, institutions, and researchers. Computers and Education: Artificial Intelligence, 1, 100005.
• Ching, F. D., & Binggeli, C. (2018). Interior design illustrated. John Wiley & Sons.
• Çelik, T. (2024). Generative design experiments with artificial intelligence: reinterpretation of shape grammar. Open House International, 49, 822–842.
• Decorative Floral Pattern. (n.d.). Retrieved from Pinterest: https://i.pinimg.com/564x/89/10/cc/8910cc027ae6616714a9e23bd79edfa4.jpg
• Decorative Pattern Design. (n.d.). Retrieved from Pinterest: https://i.pinimg.com/736x/20/5a/2f/205a2fae45b2d4b94c6c3cc054abbef8.jpg
• Dehouche, N., & Dehouche, K. (2023). What’s in a text-to-image prompt? The potential of stable diffusion in visual arts education. Heliyon, 9.
• Hanafy, N. O. (2023). Artificial intelligence's effects on design process creativity:" A study on used AI Text-to-Image in architecture". Journal of Building Engineering, 80, 107999.
• Hao, Y., Chi, Z., Dong, L., & Wei, F. (2023). Optimizing prompts for text-to-image generation. Advances in Neural Information Processing Systems, 36, 66923–66939.
• Heydarian, A., Pantazis, E., Wang, A., Gerber, D., & Becerik-Gerber, B. (2017). Towards user centered building design: Identifying end-user lighting preferences via immersive virtual environments. Automation in Construction, 81, 56–66.
• Horvath, A.-S., & Pouliou, P. (2024). AI for conceptual architecture: Reflections on designing with text-to-text, text-to-image, and image-to-image generators. Frontiers of Architectural Research, 13, 593–612.
• Ko, H.-K., Park, G., Jeon, H., Jo, J., Kim, J., & Seo, J. (2023). Large-scale text-to-image generation models for visual artists’ creative works. Proceedings of the 28th international conference on intelligent user interfaces, (s. 919–933).
• Mahdavi Goloujeh, A., Sullivan, A., & Magerko, B. (2024). Is It AI or Is It Me? Understanding Users’ Prompt Journey with Text-to-Image Generative AI Tools. Proceedings of the 2024 CHI Conference on Human Factors in Computing Systems, (s. 1–13).
• Midjourney Offical Website. (n.d.). Retrieved from www.midjourney.com
• Navarra Center Table. (n.d.). Retrieved from Boca do Lobo: https://www.bocadolobo.com/images/products/limited-edition/coffee-and-side-tables/center-table/navarra/navarra-center-table-14-zoom-boca-do-lobo.png
• Neufert, E., & Neufert, P. (2012). Architects' data. John Wiley & Sons.
• Panero, J., & Zelnik, M. (1979). Human dimension and interior space: A source book of design reference standards. Watson-Guptill.
• Parashar, B., Sharma, R., Rana, G., & Balaji, R. D. (2023). Foundation concepts for industry 4.0. In New horizons for industry 4.0 in modern business (pp. 51–68). Springer.
• Pile, J. F. (2005). A history of interior design. Laurence King Publishing.
• Product Image of PH179580. (n.d.). Retrieved from IKEA: https://image-ikea.mncdn.com/urunler/2000_2000/PH179580.jpg
• Sigov, A., Ratkin, L., Ivanov, L. A., & Xu, L. D. (2024). Emerging enabling technologies for industry 4.0 and beyond. Information Systems Frontiers, 26, 1585–1595.
• The Story Behind Mies van der Rohe's Iconic Barcelona Chair. (n.d.). Retrieved from Architectural Digest: https://www.architecturaldigest.com/story/the-story-behind-mies-van-der-rohes-iconic-barcelona-chair?epik=dj0yJnU9STNHUFhoRHpPUEtpNXlPNURya1hLaFNEeGxFdWxKWVgmcD0wJm49YnVGNlR3WndhQXA1LVhXMGxib3hNQSZ0PUFBQUFBR2J0WWcw
• Vartiainen, H., & Tedre, M. (2023). Using artificial intelligence in craft education: crafting with text-to-image generative models. Digital Creativity, 34, 1–21.
• Wasielewski, A. (2023). “Midjourney can’t count”: Questions of representation and meaning for text-to-image generators.
• Zhou, E., & Lee, D. (2024). Generative artificial intelligence, human creativity, and art. PNAS nexus, 3, 52.