enr.png

Improving Safety with Machine Learning

What if a computer could automatically identify safety risk just by looking at the photos coming from your projects?  Smartvid.io has partnered with Engineering News-Record as well as key customers to demonstrate how techniques in machine learning can help safety experts be "in more places at once" by automatically "seeing" a set of safety risk indicators and flagging them for review. The result is lower risk for your firm. 

The case study below details how Smartvid.io partnered with ENR’s editorial staff to use computer vision to analyze all 2016 ENR Photo Contest images alongside the traditional human safety experts in a Machine Learning for Construction Safety Demonstration

CASE STUDY:
MACHINE LEARNING AND CONSTRUCTION SAFETY

See the results of applying machine learning to the ENR 2016 Photo Competition to screen for safety risk.

Get the case study >

How did we detect potential safety issues?

  • VINNIE (the Smartvid.io machine learning engine) has been programmed to look for key indicators of risk in jobsite photos, including the presence of "people" and whether or not those people were wearing "hard hats" and proper "safety-colored clothing".
  • VINNIE then sorted through all 1,080 ENR contest submissions in under 10 minutes, while the human team required over 4.5 hours.
  • VINNIE detected 446 images with people, while the human review found 414.
  • VINNIE also flagged 32 images containing personnel missing hard hats, and 106 images with workers missing safety-colored clothing. 
 
We continue to push the boundaries of what is possible on our projects. It’s exciting to test innovative technologies that could have a positive impact on the entire construction experience.
 
— Martin Leik, Regional Safety Director at Suffolk Construction
  • ENR Photo Competition 2016
    VINNIE has flagged potential for missing safety vests. Image courtesy of ENR.
  • ENR Photo Competition 2016
    VINNIE has flagged potential for missing safety vests or hard hats.  Image courtesy of ENR.
  • ENR Photo Competition 2016
    VINNIE detects people in a jobsite photo. Image courtesy of ENR.
  • ENR Photo Competition 2016
    VINNIE has flagged potential for missing safety vests. Image courtesy of ENR.
  • ENR Photo Competition 2016
    VINNIE detects people in a jobsite photo. Image courtesy of ENR.
  • ENR Photo Competition 2016
    VINNIE detects people in a jobsite photo. Image courtesy of ENR.
  • ENR Photo Competition 2016
    VINNIE has flagged potential for missing safety vests. Image courtesy of ENR.
  • ENR Photo Competition 2016
    VINNIE has flagged potential for missing safety vests. Image courtesy of ENR.