Saturday, November 21, 2026
Science
No Result
View All Result
  • Login
  • HOME
  • SCIENCE NEWS
  • CONTACT US
  • HOME
  • SCIENCE NEWS
  • CONTACT US
No Result
View All Result
Scienmag
No Result
View All Result
Home Science News Technology and Engineering

Generative AI Aligns Indoor Visualizations With Resident Preferences in Global Study

November 21, 2026
in Technology and Engineering
Denise Maddox
By Denise Maddox Scienmag Editorial Profile - Mechanical Engineering
Reading Time: 6 mins read
0
Generative AI Aligns Indoor Visualizations With Resident Preferences in Global Study

Generative AI Aligns Indoor Visualizations With Resident Preferences in Global Study

65
SHARES
587
VIEWS
Share on FacebookShare on Twitter
ADVERTISEMENT

A structured, questionnaire-driven generative artificial intelligence framework can reliably convert individual spatial design preferences into preference-aligned residential interior visualizations, according to an exploratory investigation published in the Journal of Engineering and Applied Science. Researchers Mohamed Nour, Osama Abo Eleinen, Ashraf Elmokadem, and Basma Nashaat developed and evaluated an automated translation pipeline that bridges user-reported spatial desires, eight established environmental psychology theories, and multimodal image synthesis. Across an international evaluation involving 122 participants from 25 countries, the system generated perspective renderings for living rooms, bedrooms, and home offices that consistently surpassed neutral evaluative benchmarks across four distinct perceptual dimensions. The study demonstrates that grounding prompt engineering in formal spatial theory enables generative models to produce coherent early-stage design concepts without relying on designer intuition or purely aesthetic criteria.

The built environment exerts an enduring influence on human experience, particularly given that individuals spend approximately 90 percent of their lives inside buildings. Architectural attributes such as daylighting quality, biophilic greenery, spatial layout, material textures, and color palettes have long been associated with perceived comfort and spatial preference across domestic, workplace, and healthcare environments. The psychological significance of residential design gained heightened empirical prominence during the COVID-19 pandemic, when prolonged domestic confinement exposed the adverse experiential consequences of deficient domestic spaces and prompted international health organizations to formally categorize housing quality as a structural determinant of well-being. However, despite decades of empirical research documenting these relationships, early-stage architectural briefing continues to rely primarily on designer intuition, lacking systematic and accessible mechanisms to translate personal occupant preferences directly into concrete spatial proposals.

Recent advancements in generative artificial intelligence, particularly large language model-guided multimodal architectures and diffusion models, present a viable technological pathway to address this long-standing disconnect. Unlike earlier generative adversarial networks that required massive domain-specific training datasets and suffered from limited prompt responsiveness, contemporary diffusion systems translate structured natural-language inputs into atmospherically differentiated renderings at high iteration speeds and low operational costs. Nevertheless, the authors observed that contemporary architectural applications of generative artificial intelligence have remained largely superficial, treating spatial qualities as rhetorical descriptions rather than rule-governed parameters. Without transparent mapping protocols connecting user desires to prompt variables and visual outputs, generated spaces lack cross-user reproducibility, and evaluations frequently default to measuring photorealism rather than genuine preference alignment or spatial suitability.

To establish a reproducible and transparent bridge between occupant preferences and visual synthesis, the research team constructed a fifteen-item multiple-choice questionnaire administered within Google Gemini Canvas. Instrument length was purposefully constrained to 15 items to optimize psychometric fidelity, as survey literature demonstrates that instruments exceeding 20 items induce respondent fatigue and satisficing behavior, while generative diffusion models experience degraded coherence when processing excessively lengthy descriptive prompts. The questionnaire organized spatial preferences into four hierarchical categories mirroring traditional architectural briefing workflows: Space and Occupancy, Visual Attributes, Physical Attributes, and Perceptual Quality Attributes. Mandatory selection menus presented discrete visual radio cards, entirely eliminating unstructured free-text responses to ensure full prompt consistency and systematic traceability across diverse evaluators.

The framework operationalized eight distinct theoretical frameworks from environmental psychology and architectural theory, dividing them into parameter-justifying and rule-governing roles without asserting clinical health outcomes. Three frameworks provided empirical evidence justifying direct parameter encoding: Attention Restoration Theory informed lighting quality, spatial scale, and biophilic density by identifying restorative spatial properties such as fascination and extent; Stress Reduction Theory linked calm and restorative mood selections to warm illumination and organic textures; and Biophilic Design Theory governed direct plant integration, indirect material representations, and spatial analogies. The remaining five frameworks served as compositional reasoning structures: Environmental Preference Theory regulated visual complexity and compositional coherence; Privacy Theory and Proxemics dictated zoning density and boundary definition; Place Attachment Theory grounded identity-expressive styling and color palettes; the Architecture of Rhythm ensured material continuity; and the Nine-Square Grid Framework established clear functional hierarchies across spatial layouts.

Participant recruitment followed a stratified purposive sampling strategy that enrolled 122 adults across 25 countries between November 2025 and April 2026. The sample was divided into two complementary evaluative groups: Profile A comprised 35 professionals and academics in architecture and interior design who provided theoretically grounded spatial appraisal, while Profile B comprised 87 residential occupants without design training who evaluated outputs against authentic lived experience. Demographically, 51.6 percent of the sample fell within the 25 to 34 age bracket, 66.4 percent were female, and participants were distributed across six world regions. In terms of residential typologies, 51 participants evaluated living rooms, 44 assessed bedrooms, and 27 reviewed home offices, with each sub-sample surpassing established adequacy thresholds for stable mean estimation in preference-rating research.

The automated visualization pipeline executed within the Google Gemini environment converted each participant\’s 15 categorical choices into discrete prompt values via a static, author-defined mapping table. Crucially, prompt terms were authored entirely in advance at instrument design time rather than generated dynamically by language models, preventing unprompted textual embellishments. These 15 values were concatenated into a standardized sequence: Space Typology, Scale and Occupancy, Color Palette, Design Style, Natural Lighting, Wall Materiality, Flooring Preference, Furniture Morphology, Air Quality and Sensory Comfort, Biophilic Density, Emotional Mood, Artwork Character, Storage Configuration, Window Type and View, and Ceiling Height. This prompt was submitted directly to Gemini\’s multimodal image generation pipeline, returning a single perspective rendering per user without batch generation, manual filtering, or post-hoc curation, ensuring uncompromised input-output traceability.

Perceptual evaluation occurred through a structured post-design questionnaire administered in Google Forms, measuring four core constructs on a five-point Likert scale: Atmospheric Coherence, Emotional Comfort, Expectations Alignment, and Overall Satisfaction. All multi-item constructs demonstrated acceptable to good internal consistency, with Cronbach\’s alpha reaching 0.81 for Atmospheric Coherence, 0.84 for Emotional Comfort, and 0.72 for Expectations Alignment, alongside an overall instrument alpha of 0.90. One-sample t-tests comparing construct means against the neutral scale midpoint of 3.0 confirmed statistically significant above-midpoint performance across all dimensions at a Bonferroni-corrected significance threshold of alpha equals 0.0125. The observed effect sizes were exceptionally large, ranging from Cohen\’s d of 1.20 to 2.52, confirming the exploratory feasibility of the framework as a preference-elicitation and visualization instrument.

Construct-level scores preserved a stable, theoretically predicted performance hierarchy across all evaluations: Atmospheric Coherence achieved the highest mean rating at 4.61, followed by Emotional Comfort at 4.46, Overall Satisfaction at 4.34, and Expectations Alignment at 4.03. Within Atmospheric Coherence, individual items assessing layout appropriateness, color atmosphere, and the balance between complexity and coherence scored between 4.49 and 4.69. This robust performance indicates that macro-level spatial and compositional parameters translate exceptionally well through structured prompt encoding, aligning with Environmental Preference Theory\’s premise that visual coherence functions as a universally mediated cognitive mechanism across diverse individuals.

The most diagnostically significant finding emerged within Expectations Alignment, where an internal dissociation separated general atmospheric resonance from literal mental imagery matching. Item 1, measuring atmospheric preference resonance, attained a high mean of 4.27, whereas Item 8, asking whether the visualization matched what the participant had imagined, recorded the lowest item mean and greatest dispersion in the study at 3.79 with a standard deviation of 1.02. Correlational analyses revealed that Item 8 exhibited markedly lower correlations with remaining scale items, ranging between 0.22 and 0.30, indicating that participants recognized specific visual divergences while simultaneously rating the spaces as satisfying and emotionally comfortable. The authors attributed this divergence to categorical compression inherent in multiple-choice formats, where complex internal mental imagery is compressed into discrete prompt variables and subsequently compared against an uncompressed mental expectation.

Sub-group analyses revealed remarkable stability across residential typologies and participant profiles without requiring typology-specific algorithm recalibration. Living rooms exhibited the highest overall ratings, reaching an Atmospheric Coherence mean of 4.68, which researchers associated with the shared, collective nature of living room appraisal. Bedrooms demonstrated intermediate scores with greater dispersion on expectations alignment, reflecting highly individualized restoration requirements, while home offices yielded slightly more conservative scores due to competing functional demands between cognitive productivity and domestic comfort. Comparing participant profiles, residential occupants in Profile B returned marginally higher satisfaction than design professionals in Profile A, with the largest divergence occurring in Expectations Alignment, where professionals evaluated input-to-output fidelity against rigorous disciplinary criteria. However, both groups preserved the identical construct hierarchy, reinforcing cross-population transferability.

Qualitative analysis of 83 informative participant statements provided deeper diagnostic triangulation, confirming that macro-level spatial organization and atmospheric mood were perceived as highly successful, whereas criticisms remained strictly localized to specific furnishings or artwork choices. The researchers emphasized several methodological limitations, noting that non-visual sensory modalities such as acoustics, thermal conditions, and olfaction were intentionally excluded because diffusion models operate exclusively in visual media. Furthermore, because participants evaluated designs generated from their own reported preferences, self-referential bias may have directionally inflated scores relative to independent external reviews. Consequently, the authors position the framework as an exploratory pre-design briefing and preference-elicitation tool for early schematic phases rather than a fully automated replacement for professional architectural judgment or a validated clinical intervention.

Subject of Research: Generative artificial intelligence framework for preference-driven residential interior design visualization

Article Title: A generative AI framework for preference-driven indoor visualization: an exploratory questionnaire-based approach

Article References: Nour, M., Eleinen, O. A., Elmokadem, A., & Nashaat, B. (2026). A generative AI framework for preference-driven indoor visualization: an exploratory questionnaire-based approach. Journal of Engineering and Applied Science, 73(1), Article 419. https://doi.org/10.1186/s44147-026-01278-y

Image Credits: AI Generated

DOI: 10.1186/s44147-026-01278-y

Keywords: generative artificial intelligence, architectural visualization, residential interior design, environmental psychology, preference elicitation, Google Gemini, spatial preference, atmospheric coherence, emotional comfort, expectations alignment, attention restoration theory, stress reduction theory, biophilic design, prompt engineering, living room design, bedroom visualization, home office planning

Cite Scienmag News

Denise Maddox. (November 21, 2026). Generative AI Aligns Indoor Visualizations With Resident Preferences in Global Study. Scienmag. https://scienmag.com/generative-ai-aligns-indoor-visualizations-with-resident-preferences-in-global-study/

Denise Maddox. "Generative AI Aligns Indoor Visualizations With Resident Preferences in Global Study." Scienmag, 21 November 2026, https://scienmag.com/generative-ai-aligns-indoor-visualizations-with-resident-preferences-in-global-study/. Accessed 21 November 2026.

Denise Maddox. "Generative AI Aligns Indoor Visualizations With Resident Preferences in Global Study." Scienmag. November 21, 2026. https://scienmag.com/generative-ai-aligns-indoor-visualizations-with-resident-preferences-in-global-study/

Tags: architectural visualizationatmospheric coherenceattention restoration theorybedroom visualizationbiophilic designemotional comfortenvironmental psychologyexpectations alignmentgenerative artificial intelligenceGoogle Geminihome office planningliving room designpreference elicitationprompt engineeringresidential interior designspatial preferencestress reduction theory
Share26Tweet16
Previous Post

Simulations Reveal Platelet Complexions Strengthen Aluminum Alloys Through Local Dislocation Climb

Next Post

Medical Schools in New South Wales Show Disparities in Palliative Care Training

Related Posts

Dosimetry Software Platforms Show Systematic Dose Discrepancies in Radiopharmaceutical Therapy
Medicine

Dosimetry Software Platforms Show Systematic Dose Discrepancies in Radiopharmaceutical Therapy

November 21, 2026
UNFC Framework Classifies Serbia’s Old Bor Mine Tailings as Potentially Viable Resource
Earth Science

UNFC Framework Classifies Serbia’s Old Bor Mine Tailings as Potentially Viable Resource

November 21, 2026
Compact Neural Network PuddleSIFNet Improves Road Water Detection Accuracy Across Hardware Platforms
Technology and Engineering

Compact Neural Network PuddleSIFNet Improves Road Water Detection Accuracy Across Hardware Platforms

November 21, 2026
Folding Architectures Cut Hardware and Power Overhead in Two-Dimensional Image Convolution
Technology and Engineering

Folding Architectures Cut Hardware and Power Overhead in Two-Dimensional Image Convolution

November 21, 2026
Groundwater Inundation Amplifies Seismic Hazards for Alexandria’s Ancient Underground Catacombs
Technology and Engineering

Groundwater Inundation Amplifies Seismic Hazards for Alexandria’s Ancient Underground Catacombs

November 21, 2026
Simulations Show Lead-Free Perovskite Solar Cell Architecture Could Reach 32.41% Efficiency
Technology and Engineering

Simulations Show Lead-Free Perovskite Solar Cell Architecture Could Reach 32.41% Efficiency

November 21, 2026
Next Post
Medical Schools in New South Wales Show Disparities in Palliative Care Training

Medical Schools in New South Wales Show Disparities in Palliative Care Training

  • Mothers who receive childcare support from maternal grandparents show more optimized

    Mothers who receive childcare support from maternal grandparents show more parental warmth, finds NTU Singapore study

    27656 shares
    Share 11059 Tweet 6912
  • University of Seville Breaks 120-Year-Old Mystery, Revises a Key Einstein Concept

    1061 shares
    Share 424 Tweet 265
  • Bee body mass, pathogens and local climate influence heat tolerance

    682 shares
    Share 273 Tweet 171
  • Researchers record first-ever images and data of a shark experiencing a boat strike

    546 shares
    Share 218 Tweet 137
  • Groundbreaking Clinical Trial Reveals Lubiprostone Enhances Kidney Function

    531 shares
    Share 212 Tweet 133
Science

Embark on a thrilling journey of discovery with Scienmag.com—your ultimate source for cutting-edge breakthroughs. Immerse yourself in a world where curiosity knows no limits and tomorrow’s possibilities become today’s reality!

RECENT NEWS

  • Community Public Health Initiative in Italy Establishes Coordinated Youth Digital Habit Framework
  • Fermented Tilapia Formulation Boosts Umami and Acidity Using Brown Sugar and Salt
  • Broader Cardiac Phenotype Beyond Ejection Fraction Proposed for Cardiogenic Shock ECMO
  • Cancer Transformation Reactivates Repressed Retroelements Uncovering Over One Thousand Potential Biomarkers

Categories

  • Agriculture
  • Anthropology
  • Archaeology
  • Athmospheric
  • Biology
  • Biotechnology
  • Blog
  • Bussines
  • Cancer
  • Chemistry
  • Climate
  • Earth Science
  • Editorial Policy
  • Marine
  • Mathematics
  • Medicine
  • Pediatry
  • Policy
  • Psychology & Psychiatry
  • Science Education
  • Science News
  • Social Science
  • Space
  • Technology and Engineering

Subscribe to Blog via Email

Enter your email address to subscribe to this blog and receive notifications of new posts by email.

Join 5,151 other subscribers

© 2026 Scienmag - Science Magazine

Welcome Back!

Login to your account below

Forgotten Password?

Retrieve your password

Please enter your username or email address to reset your password.

Log In
No Result
View All Result
  • HOME
  • SCIENCE NEWS
  • CONTACT US

© 2026 Scienmag - Science Magazine

Discover more from Science

Subscribe now to keep reading and get access to the full archive.

Continue reading