This paper presents the development and implementation of a deep learning model for predicting and automating spatial decisions in interior design. The model combines convolutional neural networks (CNNs) for visual data analysis and recurrent neural networks (RNNs) for text query processing, which allows for comprehensive information processing and personalised design solutions. The study results show high model accuracy, low cross-entropy, and high average object and IoU accuracy, which confirms its effectiveness compared to traditional methods and existing AI models. The model's integration into an interior design decision support system is provided by a user-friendly interface and interaction with popular CAD systems, which contributes to the continuity of workflows and overall efficiency. The use of cloud-based technologies for data storage and processing ensures scalability and performance of the system. The study also identified several challenges, including ensuring data privacy and security, which require further attention. The proposed model has great potential for implementation in practice, ensuring high quality and personalisation of design solutions, taking into account the individual needs of users.
No takes yet. Share an insight, caveat, or question.
Kruk et al. (2024) studied this question.