AI-based augmented reality for architectural mass study

In the domain of architecture, the creation of a physical model of a design concept is not feasible for rapid, real-time adjustments to the model due to the lengthy and material-dependent creation process. To address this limitation, an innovative technique has been developed, utilizing image-to-image translation from artificial intelligence (AI) studies. This enables real-time visual exploration of mass studies in augmented reality (AR) by augmentation with virtual materials. The present study aims to evaluate the effects of an AI-based AR system on the transformation of material appearance in mass studies. The workflow encompasses the development of synthetic data, model training, and the establishment of the AR design environment. To provide a comprehensive understanding of the AI-based AR system’s impact, we conducted a user study involving 40 designers. The experimental results indicate that the system effectively supports the manipulation of design models’ appearance to stimulate visual explorations, allowing designers to consider the geometry of the form, alternative material options for façade and spatial perception of mass on site. The proposed environment has the potential to facilitate human–computer collaboration, providing designers with alternatives, and is eco-friendly due to its reduced reliance on waste material. The discussion has been conducted with the aim of highlighting the opportunities and challenges associated with material transformation in mass study with the implementation of an AI-based AR system.