Human-centred AI for streetscape and heritage-integrity assessment: a decision-support framework for Old Sana'a, Yemen
This study presents a human-centred, AI-assisted decision-support framework that quantifies the strong correlation between street quality and heritage integrity in Old Sana'a, Yemen, using computer vision and expert validation to generate a priority matrix for targeted conservation management.
Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer
In the heart of Yemen lies Old Sana'a, a city where history is not locked away in museums but lived in the narrow, winding streets that connect towering, decorated homes. For centuries, these streets have been more than just paths for walking; they are the living veins of a heritage that depends on the continuity of the entire neighborhood, not just the survival of individual monuments. Preserving such a place requires balancing two often competing needs: keeping the historic buildings authentic and intact, while ensuring the streets remain functional, safe, and pleasant for the people who walk them every day. Too often, experts treat these two goals as separate problems, focusing either on the beauty of the architecture or the flow of daily life. But in a living city, the condition of a wall and the clarity of a path are deeply linked. Understanding how these elements interact is crucial for making smart decisions about where to spend limited resources, especially in a city facing the pressures of conflict and time.
A team of researchers has developed a new way to look at this complex relationship, combining human expertise with artificial intelligence to map the health of Sana'a's streets. They walked through 96 different street segments, capturing thousands of photographs to create a detailed visual record of the city. Instead of relying solely on human eyes to judge every inch of every street—a task that is slow and prone to inconsistency—they used a computer system trained to spot specific details. This system, acting as a tireless assistant, identified everything from the type of windows and the condition of the brickwork to the presence of modern intrusions like exposed wires or satellite dishes. However, the researchers were careful to ensure the computer did not make the final call. They designed a process where the machine's findings were reviewed and corrected by human experts who understand the deep cultural significance of the city's architecture. This partnership allowed them to create two distinct scores for every street: one measuring how well the street functions for people, and another measuring how well the historic fabric has been preserved.
The results revealed a clear and strong connection between these two scores. Streets that were well-ordered and easy to navigate tended to also be the ones where the historic buildings were in better condition. This suggests that when a street is cared for as a whole, both its utility and its heritage value tend to thrive together. However, the study also uncovered a fascinating split depending on the type of street. The narrow residential alleys, where people live, showed the highest levels of historic integrity; the old walls and traditional designs were well-preserved, but these streets often struggled with functional issues like poor lighting or blocked pathways. In contrast, the busy commercial streets were much easier to move through and visually clear, but they suffered from a loss of historic character, often covered in modern signage, air conditioning units, and altered facades. The mixed-use streets fell somewhere in the middle, balancing both qualities.
To turn these findings into action, the researchers organized the streets into four distinct groups to guide city managers. Some streets needed urgent help because they were suffering from both decay and poor function. Others were in good shape functionally but needed strict control to stop modern intrusions from ruining their historic look. A third group had beautiful, intact buildings but needed improvements to make them safer and easier to walk through. The final group was in excellent shape and only needed routine monitoring. By using this system, city planners can now see exactly where to focus their efforts. For example, a specific commercial street might be identified as needing better management of its signage and cables rather than structural repairs, while a residential alley might need new paving and lighting without touching the historic walls.
This approach does not replace the judgment of human experts; rather, it gives them a powerful tool to see the big picture quickly and consistently. The computer handles the repetitive work of spotting thousands of small details, while the humans provide the necessary context to understand what those details mean for the city's soul. The study confirms that while artificial intelligence can detect patterns and defects with impressive speed, it cannot yet understand the deep cultural meaning of a historic city on its own. The most effective path forward is a partnership where technology provides the data and humans provide the wisdom. This method offers a practical, transparent way to protect Old Sana'a, ensuring that decisions are based on clear evidence rather than guesswork, and that the city's unique character is preserved for future generations.
Drowning in papers in your field?
Get daily digests of the most novel papers matching your research keywords — with technical summaries, in your language.