Our Problem 

Soldiers in the battlefield need a way to quickly and accurately identify lab equipment in order to recognize Weapons of Mass Destruction (WMD) threats.  Our goal is to assist the DTRA in improving the current system or find a new system to accurately identify WMDs

DTRA’s Current Approach

The Defense Threat Reduction Agency currently utilizes Deep Learning Technology to automate this process. The current system consists of a mobile app with a deep learning algorithm used to take photos of unknown laboratory equipment and quickly identify them. Because the images used to train the algorithm do not translate to operational contexts due to environmental factors that obfuscate the photos, the algorithm has a current accuracy range of 30-70%

Our Mission

To assist the DTRA in improving the current system or find a new system to accurately identify WMDs. 

Our Sponsor 

Dr. John Ewing, A&AS contractor, Research & Development Integration Division, Defense
Threat Reduction Agency (DTRA)

Cost in USD to train a soldier in WMD identification

Number of interviews conducted

%

Current accuracy of DTRA's algorithm