Conference Paper
Proceedings of BSA Conference 2024: Sixth Conference of IBPSA-Italy
Predicting Daylight Preferences Using HDRI and Deep Learning
Predicting Daylight Preferences Using HDRI and Deep Learning
Dongjun Mah, Athanasios TzempelikosDOI: https://doi.org/10.13124/9788860462022_68Abstract: This paper utilizes High Dynamic Range Imaging and deep learning that utilize pixelwise information from the entire luminance distribution in the field of view to classify daylighting preferences of office workers. Generated luminance and contrast similarity maps were used for training convolutional neural network (CNN) models to classify the occupant’s visual preferences. Preference datasets for 11 individuals, collected in real offices, were used to evaluate the preference classification performance. The results showed the superiority of the luminance similarity map as a visual preference indicator compared to common static lighting parameters. Pages: 557 - 564 Paper: