TSA plans to add machine learning to carry-on baggage scans -- GCN

The Transportation Security Agency plans to incorporate machine learning into the computer tomography scanners that are starting to be used at airport security checkpoints. To advance the Accessible Property Screening Systems program, TSA is looking for researchers and industry partners to develop algorithms that could improve the automated detection of explosives and prohibited items among carry-on baggage and speed passengers through the checkpoints. Although it has been used to screen checked luggage for explosives since 2001, CT scanning is a relatively new tool for examining carry-on baggage where it would have to identify prohibited items like knives and disassembled weapons. The 3-D imaging and detection software in the CT scanners would increase the speed and accuracy of the scans, flagging the threats operators should manually check. It may eliminate the need for passengers to put their electronics and liquids in separate screening bins, TSA said prior to a June 2017 test of the technology.