Street View Imagery

This study used historical street view imagery and cyclist data from London to establish causal relationships between urban features (like vegetation, slope, and bike lanes) and cycling patterns, finding that their impacts vary by population density and providing actionable insights for urban planning.

This study introduces a novel AI-based framework to mitigate perspective biases in vehicle-captured street view imagery, successfully translating car-centric views to pedestrian and cyclist perspectives for more accurate bikeability and walkability assessments in urban planning.

I attended IATBR 2024 in Vienna, Austria this week and presented my work on the causal impacts of visual elements on cyclist activities!

This study reviewed 393 papers on how visual characteristics of the built environment influence urban perception, highlighting trends, limitations, and future research opportunities, and demonstrated the efficacy of using machine learning to semi-automate the review process.

This study introduces a cost-effective, deep learning-based method using time-series street view imagery to evaluate and monitor neighborhood environmental quality, offering a scalable tool for improving urban policy and design.

I was invited to talk about my research on street view imagery to a group of audience from government agencies and industries

The study develops a new method using deep learning to correct perspective biases in street view imagery, improving the assessment of active transportation infrastructure from the viewpoint of cyclists and pedestrians.

Can we improve the scalability of bikeability assessment with street view imagery?

Street view imagery is booming. Check how to use it for your research.

We advanced the comprehensive assessment of bikeability using street view imagery and computer vision.