Becoming rentABLE aims to help travelers with disabilities find truly accessible short-term rentals. The platform currently verifies accessibility features manually by reviewing listing photos, a process that is slow, subjective, and difficult to scale as the number of listings grows. Hosts often upload inconsistent or incomplete photos and make verification even more challenging. To address these challenges, we developed a computer vision–based review assist system that helps identify accessibility features from listing photos. Rather than replacing human judgment, the system is designed to support reviewers by highlighting relevant features and organizing results at the property level, improving both efficiency and consistency in the verification process. The system generates structured, property-level outputs with visual evidence to support faster and more reliable validation. In parallel, the project addressed upstream and downstream workflow challenges beyond the modeling component. We proposed improvements to the host upload process to encourage more complete and consistent photo submissions, along with supporting documentation to guide hosts in capturing relevant accessibility features. In addition, we implemented enhancements to the web interface and backend data structures to enable more organized data collection and long-term system scalability, which supports more reliable mapping between images and accessibility features. Our results show that the system is effective at capturing relevant accessibility features, particularly when prioritizing recall through a sensitivity-aware detection strategy to minimize missed detections. At the same time, variability in image quality and visual ambiguity introduce challenges that require human verification and a human-in-the-loop approach. The project also strengthens the overall data pipeline and improves the consistency and scalability of accessibility verification.
Mentor: Lauren Nichols
Project poster (PDF)
