URSimulator: Human-Perception-Driven Prompt Tuning for Enhanced Virtual Urban Renewal via Diffusion Models

TL;DR

This paper introduces URSimulator, a human-perception-driven prompt tuning framework that enhances virtual urban renewal visualization by iteratively editing street images to improve perceptions of safety, beauty, and liveliness, aiding urban planning.

Contribution

It presents a novel prompt tuning method combining text-driven diffusion models with human perception feedback for urban environment simulation.

Findings

Significant perception improvements: safety +17.60%, beauty +31.15%, liveliness +28.82%.

Outperforms existing methods like DiffEdit in perception enhancement.

Effective across various urban renewal scenarios.

Abstract

Tackling Urban Physical Disorder (e.g., abandoned buildings, litter, messy vegetation, graffiti) is essential, as it negatively impacts the safety, well-being, and psychological state of communities. Urban Renewal is the process of revitalizing these neglected and decayed areas within a city to improve the physical environment and quality of life for residents. Effective urban renewal efforts can transform these environments, enhancing their appeal and livability. However, current research lacks simulation tools that can quantitatively assess and visualize the impacts of renewal efforts, often relying on subjective judgments. Such tools are crucial for planning and implementing effective strategies by providing a clear visualization of potential changes and their impacts. This paper presents a novel framework addressing this gap by using human perception feedback to simulate street environment enhancement. We develop a prompt tuning approach that integrates text-driven Stable Diffusion with human perception feedback, iteratively editing local areas of street view images to better align with perceptions of beauty, liveliness, and safety. Our experiments show that this framework significantly improves perceptions of urban environments, with increases of 17.60% in safety, 31.15% in beauty, and 28.82% in liveliness. In contrast, advanced methods like DiffEdit achieve only 2.31%, 11.87%, and 15.84% improvements, respectively. We applied this framework across various virtual scenarios, including neighborhood improvement, building redevelopment, green space expansion, and community garden creation. The results demonstrate its effectiveness in simulating urban renewal, offering valuable insights for urban planning and policy-making.