Detection and tracking models, camera calibration, and field-data workflows for stronger intersection-safety insights.
Project story
The Intersection Safety Challenge focused on extracting dependable road-user movement information from complex intersection video.
My work included model training, tracking, camera calibration, field-data collection, annotation, data management, and analysis—building an end-to-end workflow rather than an isolated model.
From field collection and annotation to calibrated movement data
Road-user detection and tracking in complex intersections
Image-space observations translated into useful movement measures
Model outputs designed for stronger safety insight
My contributions
Project gallery
Resources