Thursday, March 29, 2018

What does your smile reveal?

Faces of smiling men and women
Photo credit: © Minerva Studio / Fotolia

Men’s and women’s smiles are measurably different, and can be used to identify gender. Researchers at the University of Bradford used artificial intelligence (AI) to analyze the dynamic movement of 109 participant smiles, determining gender with 86% accuracy. The testing done was relatively simple. Accuracy can improved. The team would like to see how AI responds to the smile of a transgender person, or a smile altered by cosmetic surgery. 

The team led by Professor Hassan Ugail used artificial intelligence to study dynamic movement instead of static images. They measured 49 points around the face, nose and eyes, as well as muscle movement during the smile — how much, how far, how fast. As might be expected, women tend to smile more broadly then men.

The emphasis of the research was more about machine learning capability, but researchers believe the dynamic smile of an individual may be unique enough to one day become a next-generation biometric identifier.


Is your smile male or female? Mapping the dynamics of a smile to enable gender recognition.

Published by the University of Bradford, Mar., 14, 2018. Retrieved Mar. 27, 2018, Science Daily.

Accurate Biometrics

Practical solutions for fingerprint collection and processing

Thursday, March 15, 2018

Biometric ID in Ancient Times?

Photo: Examples of ancient biometrics

We see in prehistoric man attempts to mark a place as one's own, to be remembered, to create something beautiful. Early cave paintings dating back around 40,000 years show human hands stenciled with red and brown pigments. Interestingly, scientists studying the handprints for gender differences find the handprints to be mostly female.

The first recorded evidence of writing is an intricate combinations of lines — called cuneiform — in clay tablets circa 3200 BC from Sumer, a region in what is now Iraq. In a developing society where most people could not read or write, a signature or identifier was needed for official documents. Thus, we see evidence circa 2600 to 2350 BC of fingerprints on clay seals. There is also evidence that ancient Babylonians used fingerprints on contracts circa 1900 BC.

Ancient China had a vast, growing population and needed a useful method of record-keeping to manage society, and keep law and order. From the Qin and Han Dynasties through the Six Dynasties periods (c206 BC to 589 AE), fingerprints were found recorded on a wide array of official documents: engagements, divorces, deeds, records of indenture and army records. They are also found used as a signature on formal confessions — not so different from modern society.

The use of biometrics as identifiers today is often seen as an encroachment of technology on modern life. From a fingerprint in clay in Mesopotamia to a fingerprint on silk in ancient China, the reality is that biometrics were used as identifiers thousands of years ago. Ancients understood that physical attributes could be used to minutely differentiate between people and used the information to great advantage.



Biometrics in the Ancient World: The Secret History of Identity

Crossmatch blog, Crossmatch.com


Accurate Biometrics

Practical solutions for fingerprint collection and processing

Tuesday, March 6, 2018

The Fascinating History of Face ID

Photo contrasting early and modern facial recognition
Photo illustration by Slate.  

For most of human history, face-to-face interactions formed the backbone of society, and faces were relied upon as early standards of identification and trust. Population growth and the urbanization of the Industrial Revolution disrupted the reliability of knowing every face in a small community. Alternatives to recognizing and authenticating individuals were needed.

A primitive first step in adopting the face as a biometric identifier was the written description. Physical descriptions sought to be objective — describing face, hair and eye color; face, nose and lips shape; and identifying facial marks and scars. While better than nothing, the system was inexact and not tremendously helpful.

The next advancement was the to represent the face visually with hand drawn sketches. Better. But it wasn't until the invention and widespread use of photography that capturing facial images became really useful. In society, cartes de visite were all the rage — playing card size photos in sepia tone that were exchanged with family and friends. For police, mug shots were a revolutionary tool to find and apprehend suspects. Criminologists studied mug shots as a tool to advance theories in linking certain physical traits to deviant behavior.

In the late 19th century, a new technique called bertillonage was developed by French policeman Alphonse Bertillon as a promising, standardized biometric identification system. The system involved the meticulous measurement of 11 parts of the body, including the head and face. These measurements, along with mug shots were considered the high tech classification system of its time.

In practice, bertillonage was a nightmare, as prisoners were not cooperative about standing in complicated poses while precise measurements were taken. Fingerprinting began its rise as the gold standard for biometric identification, while other traits such as voice, iris, gait and genetics drew interest and study.

Advances in facial recognition continued. Computers were used to study specific facial markers in the 1960's and 1970's. By the 1990's digital facial software was used by state departments of motor vehicles to prevent ID theft and fraud in driver's licenses.

The 9/11 attacks and the "War on Terror" brought facial recognition back to the forefront as the study of mass surveillance tactics vastly expanded. The FBI developed its Next Generation Identification System. Described as "the world's largest and most efficient electronic repository of biometric and criminal history information", it provided law enforcement a search and response system that included facial recognition in 2014. 

The Department of Homeland Security turned to face recognition for border security and is now scanning international travelers to make sure the identity of the traveler matches the photo on the passport. Other countries are now using facial recognition for border protection as well. 

Private companies joined the race for a highly accurate and secure authentication system. Fingerprint sensors became widespread for mobile devices and business security. Then last year Apple introduced the iPhone X with a facial recognition systems that's secure enough to determine the difference between identical twins.

Artificial software gets trained on the massive amount of photos uploaded to the internet daily, with applications such as Facebook's DeepFace. We have fears about technology getting away from us. Alvaro Bedoya, executive director of Georgetown Law's Center on Privacy and Technology stated, "I know what I touch, and I certainly know if I give fingerprints for a background check. I don't think there's anyone who keeps track of every surveillance or smartphone camera."

It's clear that we we need to be mindful of the information and photos we're sharing online. As technology develops at rocket speed, we need to rethink privacy policies, and set reasonable limits, involving government to protect us from potential misuse. "For without reasonable limits, these systems that attempt to capture and authenticate a version of our bodies may belie our belief that we own them."


What's in a Face ID?

By José Ragas, Mar. 5, 2018, for Slate.com


Accurate Biometrics

Practical solutions for fingerprint collection and processing