Classifying Cycling Hazards in Egocentric Data

 This project is an egocentric video data-set of hazardous cycling situations. The database contains short videos of some hazardous cycling situations recorded by a cyclist. Along with these videos, there are IMU measurements and classifications of the hazard completed with Amazon Mechanical Turk . The resulting database will be used to analyse cycling infrastructure and help its improvement.

Jayson Haebich
Senior Researcher