The second half of the 20th century witnessed great bounds in terms of the capabilities of biometric technology. As the number of documented people grew, a need for increased automation and efficiency in identification became readily apparent. These issues, paired with more prospects of the potential applications of biometrics motivated increased efforts to innovate in this area, especially in the governmental sphere. In 1964, Woodrow W. Bledsoe, Helen Chan Wolf, and Charles Bisson pioneered semi-automated facial recognition technology (FRT) during their studies of pattern recognition intelligence. The U.S. government subsequently contracted Bledsoe to take this system further in order to utilize it for security purposes (West, 2017). As a result, a system was developed that used manually selected points of focus on a photograph of a face (e.g. eyes, nose, mouth, ears), that the system then took measurements and ratios of in reference to a single point in order to create a 1D map that could be compared to previously collected reference data. This initial breakthrough was carried further by Goldstein, Harmon, and Lesk in the 1970s when they implemented a system of 21 subjective markers (e.g. lip thickness, hair color) to increase the specificity and automation of FRT; however, the process still required an administrator to manually input these data points into the system. The next milestone was reached by Sirovich and Kirby, who realized in 1988 that principles of linear algebra could be used to code a representative facial image in what they called the Eigenfaces technique. They found that less than one hundred values were required to create a basic framework of facial features that could be analyzed, an idea Turk and Pentland expanded on to make automated and reliable real-time FRT a reality by using the residual error from the Eigenfaces technique in 1991 (Turk, 1991). As innovation continued and the applications of FRT became more feasible, government agencies sought to promote the creation of a large enough database of registered faces in order to spur advancement into the commercial sector. This led to the development of the FacE REcognition and Technology (FERET) Program in 1993 by the Defense Advanced Research Projects Agency (DARPA) and the National Institute of Standards and Technology (NIST). In the early 2000s, Face Recognition Vendor Tests (FVRT) were implemented to build on FERET and establish performance evaluations of commercialized FRT by the federal government in order to better direct its usage by law enforcement
11 and government agencies. It was determined that the most effective use of FRT by law enforcement agencies was to capture images of perpetrators in real-time through the use of body cameras and run comparison tests to the database contained within the state’s Department of Highway Safety and Motor Vehicles. This system was implemented in 2009 by the Pinellas County Sheriff’s Office and helped facilitate many arrests and investigations into criminal activity that would not have been possible otherwise by 2011.