Face Recognition Algorithms Showing Significant Improvement, NIST Says
Face recognition algorithms used in testing in large population sets have demonstrated “massive gains in accuracy” over the past five years, according to a new report by the National Institute of Standards and Technology (NIST). The November 2018 report, Ongoing Face Recognition Vendor Test (FRVT) Part 2: Identification, tested 127 algorithms from 45 developers against a database of 26.6 million live portrait photos of 12.3 million individuals. The testing also involved three smaller datasets against more unconstrained photos. “The major result of the evaluation is that massive gains in accuracy have been achieved in the last five years (2013-2018) and these far exceed improvements made in the prior period (2010-2013),” the report says. “While the industry gains are broad—at least 28 developers’ algorithms now outperform the most accurate algorithm from late 2013—there remains a wide range of capabilities. With good quality portrait photos, the most accurate algorithms will find matching entries, when present, in galleries containing 12 million individuals, with error rates below 0.2%.” Things like aging, injury and substandard image quality lead to higher error rates, requiring human adjudication, it says. The gains in face recognition algorithm accuracy are due to technology improvements, in particular, “deep convolutional neural networks,” NIST says. “As such, face recognition has undergone an industrial revolution, with algorithms increasingly tolerant of poor quality images.”
MSA Security Passes Testing for TSA Canine Air Cargo Program
MSA Security says that that its first 10 Windsor canine bomb-sniffing teams have passed testing overseen by the Transportation Security Administration in anticipation of the start of the third-party canine explosive detection canine program for air cargo. The program is part of a federal directive requiring all cargo shipped in passenger aircraft be screened for explosives. Following the initial round of testing, MSA and other third-party providers of explosive detection canine teams can apply to become Certified Cargo Screening Facility-K9 providers. The K9 providers can be contracted by all cargo carriers, airlines, freight forwarders, and other TSA-regulated parties to deploy explosive detection canine teams for safe and efficient cargo screening.
Jet Blue Introduces Biometric Self-Boarding Gate at JFK
JetBlue earlier this month rolled out its first fully-integrated biometric self-boarding gate at John F. Kennedy International Airport in New York for customers flying to select international destinations from Terminal 5. The dual lane gate uses facial recognition technology to verify travelers. The new self-boarding gate follows successful biometric trials by the airline at Boston Logan International, Fort Lauderdale-Hollywood International in Florida, and JFK. JetBlue says it plans to expand the biometric boarding program to more international flights from JFK, Boston and Fort Lauderdale, and expects to pilot a biometric bag drop station at JFK early in 2019. The airline is working with Customs and Border Protection on the deployments. CBP uses the facial recognition technology to verify that foreign nationals are departing the U.S. in accordance with their visa terms. Separately, the Metropolitan Washington Airports Authority, in partnership with JetBlue, has launched a one-step biometric boarding process at Reagan National Airport using the authority’s veriScan facial recognition system. The veriScan system allows passengers to board a JetBlue flight from Reagan to Nassau in the Bahamas without presenting a passport or boarding pass. MWAA says the system was launched on Nov. 10, saying the airline is the first to deploy veriScan in a one-step allowing passenger boarding with just the facial recognition technology. It says the system allowed JetBlue to board 93 passengers in 10 minutes while enabling verification by both CBP and the airline’s departure control system. The deployment at Reagan helps CBP meet a congressional mandate to implement biometric exit processing for all commercial flights leaving the U.S.