• June 12, 2017

    PacTrans Partnership with City of Bellevue and Microsoft

    In 2016, road crashes resulted in 40,000 deaths and 4.6 million injuries in the United States. The FHA recently reported that more than 50 percent of the combined total of fatal and injury crashes occur at or near intersections. These numbers may sound large, but for any given location, this number is actually quite small. Thus, it takes several years to gather enough crash data to identify a problem area, then several more years to plan, fund, and implement a fix. Meanwhile, exponentially more near-accidents, and small accidents that go unreported, happen every day at those same locations.

    The ability to leverage these near accidents, as well as the small accidents that would otherwise go unreported, can provide significantly more information from which to better inform decision makers on where to invest scarce resources to make the greatest impact to the safety of our transportation networks. Thus the City of Bellevue has partnered with the University of Washington and Microsoft to develop a program that utilizes already-in-operation traffic cameras and machine learning, can detect these other events in real time. This will fundamentally change the way safety alterations are made from a reactive process to a proactive one.

    PacTrans has written on this partnership several times in the past. Bellevue Principal Transportation Planner, Franz Loewenherz, recently held a press conference to launch the crowdsourcing portion of this project where we invite the general public to visit the website, and help us train the digital neural network that is behind the machine learning method to this program. This press conference has generated significant interest in the press, including: on Geekwire, MetroLab Network, Institute of Transportation Engineers, Mobility Lab, Bellevue Patch, Seattle Transit Blog, KOMO-AM (Radio), KOMO-SEA 1 (ABC), KOMO-SEA 2 (ABC), and KIRO 7 TV.

    Please follow the link above to the project website, learn more about the partnership, and help us label and track the movement of people in the short clips of pre-recorded video footage!