Enhancing visual comfort prediction in space layouts using deep learning models

ABSTRACT

In designing performance-based facilities, such as educational buildings, assessing visual comfort is crucial. The computational cost and resource-intensive modeling necessitate an efficient analytical tool for diverse layout assessment. In the current study, predictive models were developed using deep learning and machine learning to predict visual comfort in typical elementary schools in Tehran. Firstly, layouts were modeled, considering influential parameters. Subsequently, simulation-based datasets were analyzed and labeled. The VGG16 and VGG19 Architectures from convolutional neural networks (CNNs), along with the pix2pix model from generative adversarial networks for forecasting respectively, numerical and pictorial indices regarding each visual comfort metric. The pix2pix model performs approximately SSIM of 0.9 for sDA, ASE, and UDI_exceed3000 and 0.86 for DAp. The extracted features by CNNs were harnessed in training models. Eventually, models with the Bayesian Ridge algorithm had a promising performance which exhibited acceptable R² values for metrics in the range of 0.90–0.96. Toward a depth analysis, parameter sensitivity using the Shapley additive explanations method was evaluated with XGBoost models. Additionally, the Spearman Correlation underscores the substantial impact of factors like WWR and aspect ratio on metrics.

GRAPHICAL ABSTRACT

KEYWORDS:

• Visual comfort
• daylight metrics
• deep learning
• decision-making framework
• early-phase assessment

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

Raw data were generated at Shahid Beheshti University. Derived data supporting the findings of this study are available from the corresponding author Amirali Ghourkhanehchi Zirak on request.

Additional information

Notes on contributors

Amirali Ghourkhanehchi Zirak

Amirali Ghourkhanehchi Zirak is a building science researcher with a master's degree in Architecture and Energy from SBU. His research focuses on Indoor Environment Quality (IEQ) Assessment, Building Performance Modeling, and Machine Learning.

Afshin Saeedian

Afshin Saeedian is a building science researcher with expertise in the implementation of artificial intelligence in built environments, particularly in the realm of deep learning.

Zahra Sadat Zomorodian

Zahra (Mahsa) Zomorodian is an Assistant Professor at the SBU Department of Construction. Her research expertise is in Thermal Comfort, Daylighting, Building Energy Simulation (BES), Indoor Environmental Quality (IEQ) assessments and Integrated Green Design Process (IGDP). She is the head leading of the Building physics lab in SBU, an inter-disciplinary group with a grounding in architecture that develops design workflows, tools and metrics to evaluate the environmental performance of buildings.

Mohammad Tahsildoost

Mohammad Tahsildoost is an Associate Professor at the SBU Department of Construction.