AI-driven architectural document parsing

Victoria Suen, School of Engineering & Technology 

This study explored the potential of AI to automate architectural document reviews, aiming to reduce delays in the city approvals process and support Ontario’s goal of building 1.5 million homes over the next decade. The project focused on developing tools to extract and analyze floor plan data, identifying key structural components such as walls, windows, doors, and room boundaries. The system also validated room areas against Ontario Building Code requirements, ensuring compliance and providing insights for potential redesigns. 

The research demonstrated the feasibility of automating aspects of architectural document review, laying the groundwork for further advancements in AI-driven tools for the construction and design industries. Next steps include expanding the tool’s capabilities, integrating machine learning for improved object recognition, and refining the user interface to enhance accessibility for architects and developers. The research team plans to share findings with faculty and industry professionals, with potential funding applications to further develop and implement these AI solutions in real-world housing projects. 

This project draws on research supported by the Social Sciences and Humanities Research Council.