The Evolution of Biometric Devices - Part One
Biometric devices measure human features in order to carry out tasks, one of the most prevalent examples being the facial recognition utilised to unlock smartphones. This is the expectation of modern handsets, and has no doubt vastly improved the security of devices and the data therein. In this article we discuss a brief chronology of the origins of biometrics and how it found its way into our very hands.
Biometrics is the technical term for body measurements, or more specifically the measurements of distinctive characteristics present in the human body. Although we are discussing the technical applications, biometrics is something that has been always been used by humans, as the very process by which we recognise one another.
As a science, biometry was first embraced as early as 1858, where Sir William Herschel, while working in India, recorded a handprint on the back of employment contracts to distinguish between true employees, and those claiming to be employees come payday. This was the first time hand and fingerprints were used to distinguish between people in such a way. In 1870, Alphonse Bertillon, a French police officer, developed anthropometries which were a way of identifying individuals through detailed records of their body measurements and photographs. He developed his system as a way to detect those offenders who gave false aliases when arrested, and it turned out to be a fairly robust system. Police forces continued to use this system alongside fingerprint records until 1903, when the Bertillon system failed after a well-documented court case: two men, who were subsequently determined to be identical twins, were accused of murder, and under the Bertillon system it was virtually impossible to differentiate the two men from one another and therefore ascertain who was the guilty party.
In 1936 ophthalmologist Frank Burch proposed the concept of using a person’s iris as a method of recognition, believing no two were the same, like fingerprints. Although sound in theory, there were no further advances on this hypothesis for quite some time.
In 1969, the FBI began a drive to develop a system to automate its fingerprint recognition system, which had previously been a manual process. They contacted the National Institute of Standards and Technology and asked them to study and evaluate the concept with a view to a computerised solution. Not so long after in 1970, Goldstein, Harmon, and Lesk, who were forensic pathologists, began developing a system to automate facial recognition. Their system used 21 markers such as hair colour and lip thickness. While initially successful, it was laborious as the initial calculations were done manually.
In 1974 biometric technology was consolidated, with the first hand geometry system becoming commercially available on the open market. Although large and cumbersome, this system was very successful and was implemented for three main purposes. The first being physical access control, the second being time and attendance, and the third being personal identification. This was arguably the device which jet propelled the research and development in its field, and in the next article we will track the onwards progress to the modern day.
In 1975 as part of its ongoing research and development scheme, the FBI funded scanners that were even more adept at scanning and charting minute biometry, specifically fingerprints. Only these minute details were stored, as opposed to the entire fingerprint, due to the lack of image compression techniques, but this spurred the acknowledgment that the minute details logged were sufficient to identify individual fingerprints, without having the complete image.
On a slight deviation, in 1977 a patent was awarded for the acquisition of signature characteristics via a digital reader. This was specifically directed towards the considerably more niche area of handwriting recognition, but was nonetheless a highly important development that has many practical applications (albeit in different iterations) today.
In 1985, a patent was awarded to Joseph Rice that was specific to ‘vascular recognition’. Vascular recognition is the process whereby an individual can be identified via their vein patterns. Also known as vascular pattern recognition, in the modern day this technology has already been used in Japan in ATM machines, as a completely contactless was of identifying the user to authenticate a transaction.