Tag Archives: biometric product marketing expert

DNA mixture interpretation outside of the forensic laboratory? Apparently not yet.

(Part of the biometric product marketing expert series)

The National Institute of Standards and Technology has published a draft report entitled DNA Mixture Interpretation: A Scientific Foundation Review.

As NIST explains:

This report, currently published in draft form, reviews the methods that forensic laboratories use to interpret evidence containing a mixture of DNA from two or more people.

From https://www.nist.gov/dna-mixture-interpretation-nist-scientific-foundation-review

The problem of mixtures is more pronounced in DNA analysis than in analysis of other biometrics. You aren’t going to encounter two overlapping irises or two overlapping faces in the real world. (Well, not normally.)

By Olli Niemitalo – Own work, CC0, https://commons.wikimedia.org/w/index.php?curid=18707318

You can certainly encounter overlapping voices (in a recorded conversation) or overlapping fingerprints (when two or more people touched the same item).

From https://fingerprintindividualization.blogspot.com/2016/07/pattern-interference-in-fingerprint.html

But there are methods to separate one biometric sample from another.

It’s a little more complicated when you’re dealing with DNA.

Distinguishing one person’s DNA from another in these mixtures, estimating how many individuals contributed DNA, determining whether the DNA is even relevant or is from contamination, or whether there is a trace amount of suspect or victim DNA make DNA mixture interpretation inherently more challenging than examining single-source samples. These issues, if not properly considered and communicated, can lead to misunderstandings regarding the strength and relevance of the DNA evidence in a case.

From the Abstract in https://doi.org/10.6028/NIST.IR.8351-draft%C2%A0

As some of you know, I have experience with “rapid DNA” instruments that provide a mostly-automated way to analyze DNA samples. Because these instruments are mostly automated and designed for use by non-scientific personnel, they are not able to analyze all of the types of DNA that would be analyzed by a forensic laboratory.

Therefore, this draft document is silent on the topic of rapid DNA, despite the fact that co-author Peter Vallone has years of experience in rapid DNA.

I am not a scientist, but in my view the absence of any reference to rapid DNA strongly suggests that it’s premature at this time to apply these instruments to DNA mixtures, such as rape cases in which both the assailant’s and the victim’s DNA are present in a sample.

Granted, there may be rape cases in which the DNA of the assailant may be present with no mixture.

You have to be REALLY careful before claiming that rapid DNA instruments can be used to wipe out the backlog of rape test kits. However, rapid DNA can be used to clear less complicated DNA cases so that the laboratories can concentrate on the more complex cases.

Putting your finger on the distribution of latent prints (the 30% palm estimate)

(Part of the biometric product marketing expert series)

Back when automated fingerprint identification systems (AFIS) were originally expanded to become automated fingerprint/palmprint identification systems (AFPIS), a common rationale for the expansion was the large number of unsolved latent palmprints at crime scenes.

By Etan J. Tal – Own work, CC BY-SA 3.0, https://commons.wikimedia.org/w/index.php?curid=41152228

The statistic that everyone cited was a statistic that 30% of all latent friction ridge prints at crime scenes were from palmprints. Here’s a citation from the National Institute of Justice.

Anecdotally, it is estimated that approximately 30% of comparison cases involve palm impressions.

Note that the NIJ took care to include the word “anecdotally.” Others don’t.

It is estimated that 30 percent of latent prints found at crime scenes come from palms.

But who provided the initial “30% of latents are palms” estimate long ago? And what was the basis for this estimate? This critical information seems to have been lost.

By Apneet Jolly – originally posted to Flickr as Candy corn contest jar, CC BY 2.0, https://commons.wikimedia.org/w/index.php?curid=10317287

Now I don’t have a problem with imprecise estimates, provided that the assumptions that go behind the estimate are well-documented. I’ve done this many times myself.

But sadly, any assumptions for the “30% of latents are palms” figure have disappeared over the years, and only the percentage remains.

Is there any contemporary evidence that can be used to check the 30% estimate?

Yes.

The blind proficiency study wasn’t blind regarding the test data

Latent print quality in blind proficiency testing: Using quality metrics to examine laboratory performance. https://lib.dr.iastate.edu/csafe_pubs/84/

A Center for Statistics and Applications in Forensic Science study (downloadable here) was published earlier this year. Although the study was devoted to another purpose, it touched upon this particular issue.

The “Latent print quality in blind proficiency testing: Using quality metrics to examine laboratory performance” study obviously needed some data, so it analyzed a set of latent prints examined by the Houston Forensic Science Center (HFSC) over a multi-year period.

In the winter of 2017, HFSC implemented a blind quality control
program in latent print comparison. Since its implementation, the
Quality Division within the laboratory has developed and inserted
290 blind cases/requests for analysis into the latent print comparison unit as of August 4, 2020….

Of the 290 blind cases inserted into casework, we were able to
obtain print images for 144 cases, with report dates spanning approximately two years (i.e., January 9, 2018 to January 8, 2020)….

In total, examiners reviewed 376 latent prints submitted as part
of the 144 blind cases/requests for analysis.

So, out of those 376 latent prints, how many were from palms?

The majority of latent prints were fingerprints (94.3%;
n = 350) or palm prints (4.9%; n = 18). Very few were joint impressions or unspecified impressions (0.8%; n = 3)….

The remaining 5 of 376 prints were not attributed to an anatomical source because examiners determined them to be of no comparative value and did not consider them to be latent prints.

For those who are math-challenged, 5 percent is not equal to 30 percent. In fact, 5 percent is much less than 30 percent. (And 4.9% is even less, if you want to get precise about it.)

Now I’ll grant that this is just one study, and other latent examinations may have wildly different percentages. At a minimum, though, this data should cause us to question the universally-accepted “30%” figure.

As any scientific institute that desires funding would proclaim, further research is needed.

And I’ll grant that. Well, I won’t grant it, but some government or private funding entity might.

How the “CSI effect” can obscure the limited role of DNA-based investigative leads

(Part of the biometric product marketing expert series)

People have been talking about the “CSI effect” for decades.

In short, the “CSI effect” is characterized as the common impression that forensic technologies can solve crimes (and must be used to solve crimes) in less than an hour, or within the time of a one-hour television show.

Fair use, https://en.wikipedia.org/w/index.php?curid=13587416

When taken to its extreme, juries may ask why the law enforcement agency didn’t use advanced technological tools to solve that jaywalking case.

Advanced technological tools like DNA, which has been commonly perceived to be the tool that can solve every single crime.

Well, that and video, because video is powerful enough to secure a conviction. But that’s another story.

Can DNA result in an arrest in a Denver homicide case?

A case in point is this story from KDVR entitled “DNA in murder case sits in Denver crime lab for 11 months.”

This is a simple statement of fact, and is not that surprising a statement of fact. Many crime labs are inundated with backlogs of DNA evidence and other forensic evidence that has yet to be tested. And these backlogs ARE creating difficulties in solving crimes such as rapes.

But when you read the article itself, the simple statement of fact is painted as an abrogation of responsibility on the part of law enforcement.

A father is making an emotional plea and putting up $25,000 of his own money to help find his son’s killer.

He is also asking the Problem Solvers to look into the time it has taken for DNA evidence to be tested in this case and others.

Tom O’Keefe said it’s taking too long to get answers and justice.

From this and other statements in the article, a picture emerges of an unsolved crime that can only be solved by the magical tool of DNA. If DNA is applied to this, just like they do on TV, arrests will be made and the killer will be convicted.

So why is it taking so long to do this?

Why is justice not being served?

KDVR is apparently not run by impassioned activists, but by journalists. And it is important from a journalistic perspective to get all sides of the story. Therefore, KDVR contacted the Denver Police Department for its side of the story.

The Denver Police Department has identified all parties involved, and the investigation shows multiple handguns were fired during this incident. While this complex case remains open, which limits details we can provide, we can verify that a significant amount of forensic work has been completed, but some remains. Investigators believe the pending forensic analysis can potentially support a weapon-related charge but will not further the ongoing homicide investigation.

OK, let’s grant that they’re not trying to identify an unknown assailant, since “all parties involved” are known.

But once that DNA is tested, isn’t that going to be the magic tool that provides the police with probable cause to arrest the killer?

Um, no.

Even IF the DNA evidence DOES happen to show a significant probability that an identifiable person committed the homicide, that in itself is not sufficient reason to arrest someone.

Why not?

Because you can’t arrest someone on DNA evidence alone.

DNA evidence can provide an investigative lead, but it has to be corroborated with other evidence in order to secure an arrest and a conviction. (Don’t forget that the evidence has to result in a conviction, and in most of the United States that requires that the evidence show beyond a reasonable doubt that the person committed the crime.)

Why was a serial killer in three European countries never brought to justice, despite overwhelming DNA evidence?

Reasonable schmeasonable.

If DNA ties someone to a crime, then the person committed the crime, right?

Let’s look at the story of a serial killer who terrorized Europe for over a decade, even though ample DNA evidence was found at each of the murder scenes, beginning with this one:

In 1993, a 62-year-old woman was found dead in her house in the town of Idar-Oberstein, strangled by wire taken from a bouquet of flowers discovered near her body.

Nobody had any information on what might have happened to Lieselotte Schlenger. No witnesses, no suspects, no signs of suspicious activity (except for the fact that she’d been strangled to death with a piece of wire, of course). But on a bright teacup near Schlenger, the police found DNA, the only clue to surface at all.

The case went cold, given that the only lead was the DNA of an unknown woman, and there was no match. Yet.

Eight years later, in 2001, there was a match when the same woman’s DNA was found at a murder scene of a strangulation victim in Freiburg, Germany. Police now knew that they were dealing with a serial killer.

But this time, the woman didn’t wait another eight years to strike again.

Five months after the second murder scene, her DNA showed up on a discarded heroin syringe, after a 7-year-old had stepped on it in a playground in Gerolstein. A few weeks later it showed up on an abandoned cookie in a burgled caravan near Bad Kreuznach, like she’d deliberately spat out a Jammy Dodger as a calling card. It was found in a break-in in an office in Dietzenbach, in an abandoned stolen car in Heilbronn, and on two beer bottles and a glass of wine in a burgled bar in Karlsruhe, like she’d robbed the place but stuck around for a few cheeky pints.

And her activities were not confined to Germany.

Over the apparent crime spree, her DNA was sprayed across an impressive 40 crime scenes in Austria, southern Germany, and France, including robberies, armed robberies, and murders.

In 2009, the case took an even more bizarre turn.

Police in France had discovered the burned body of a man, believed to be from an asylum seeker who went missing in 2002. During his application, the man had submitted fingerprints, which the police used to try and confirm his identity. Only, once again, they found the DNA of the phantom.

“Obviously that was impossible, as the asylum seeker was a man and the Phantom’s DNA belonged to a woman,” a spokesperson for the Saarbrücken public prosecutor’s office told Spiegel Online in 2009.

But how could this be?

DNA evidence had tied the woman, or man, or whatever, to six murders and numerous other crimes. There was plenty of evidence to identify the criminal.

What went wrong?

Well, in 2009 police finally figured out how DNA evidence had ended up at all of these crime scenes in three countries.

The man’s death led to an explanation of the case: there was no serial killer, and the DNA could be traced to a woman working in a packing center specializing in medical supplies. It was all down to DNA contamination.

Well, couldn’t that packing woman be convicted of the serial murders and other crimes, based upon the DNA evidence?

No, because there was no other evidence linking the woman to the crimes, and certainly “reasonable doubt” (or the European criminal justice equivalent) that the woman was also the dead male asylum seeker.

This is why DNA is only an investigative lead, and not evidence in and of itself.

But the Innocence Project always believes that DNA is authoritative evidence, right?

Even those who champion the use of DNA admit this.

If you look through the files of people exonerated by the Innocence Project, you find a common thread in many of them.

Much of the evidence gathered before the suspect’s original conviction indicated that the suspect was NOT the person who committed the crime. Maybe the family members testified that the suspect was at home the entire time and couldn’t have committed the crime in question. Or maybe the suspect was in another city.

However, some piece of evidence was so powerful that the person was convicted anyway. Perhaps it was eyewitness testimony, or perhaps something else, but in the end the suspect was convicted.

Eventually the Innocence Project got involved, and subsequent DNA testing indicated that the suspect was NOT the person who committed the crime.

This in and of itself didn’t PROVE that the person was innocent, but the DNA test aligned with much of the other evidence that had previously been collected. It was enough to cast a reasonable doubt on the conviction, allowing the improperly convicted suspect to go free.

But there are some cases in which the Innocence Project says that even DNA evidence is not to be trusted.

Negligence in the Baltimore Police Department’s crime lab tainted DNA analysis in an unknown number of criminal cases for seven years and raises serious questions about other forensic work in the lab, the Innocence Project said today in a formal allegation that the state is legally required to investigate.

DNA contamination, the same thing that caused the issues in Europe, also caused issues in Baltimore.

And there may be other explanations for how a person’s DNA ended up at a crime scene. Perhaps a police officer was careless and left his or her DNA at a crime scene. Perhaps someone was at a crime scene and left DNA evidence, even though that person had nothing to do with the crime.

In short, a high probability DNA match, in and of itself, proves nothing.

Investigative leads and reasonable doubt are very important considerations, even if they don’t fit into a one-hour TV show script.

Investigative leads and DNA booking stations

(Part of the biometric product marketing expert series)

From https://www.facebook.com/groups/bredemarketidentity/posts/756204735074350

A July Bredemarket post on Facebook has garnered some attention in September.

I wanted to answer some questions about rapid DNA use in a booking station, how (and when) DNA is used in booking (arrests), what an “investigative lead” is, and whether acquiring DNA at booking is Constitutional.

(TL;DR on the last question is “yes,” per Maryland v. King.)

Are rapid DNA booking stations a Big Brother plot?

The post in question was a Facebook post to the Bredemarket Identity Firm Services Facebook group. I posted this way back in July, when Thermo Fisher Scientific became the second rapid DNA vendor (of two rapid DNA vendors; ANDE is the other) whose system was approved by the U.S. Federal Bureau of Investigation (FBI) for use as a law enforcement booking station.

When I shared this on Facebook, I received some concerned comments:

“Big brother total control”

“Is this Constitutional??? Will the results of this test hold up in courtrooms???”

I’ll address the second question later: not just in regard to rapid DNA, but to DNA in general. At this point, however, I will go ahead and say that the use of rapid DNA in booking was authorized legislatively by the Rapid DNA Act of 2017. This was followed by over three years of procedural stuff until rapid DNA booking station use was authorized this year.

To accurately state what “rapid DNA booking station use” actually means, let me refer to the FBI’s language, starting with the purpose:

The FBI Laboratory Division has been working with the FBI Criminal Justice Information Services (CJIS) Division and the CJIS Advisory Policy Board (CJIS APB) Rapid DNA Task Force to plan the effective integration of Rapid DNA into the booking station process.

By way of definition, a “booking station” is a computer that processes individuals who are “booked,” or arrested. The FBI’s plan was that (when authorized by federal, state, or local law) when an arrested individual’s fingerprints were captured, the individual’s DNA would be captured at the same time. (Again, only when authorized.)

The use of the term “reference sample buccal (cheek) swab” is intentional. The FBI’s current development and validation efforts have been focused on the DNA samples obtained from known individuals (e.g., persons under arrest). Because known reference samples are taken directly from the individual, they contain sufficient amounts of DNA, and there are no mixed DNA profiles that would require a scientist to interpret them. For purposes of uploading or searching CODIS, Rapid DNA systems are not authorized for use on crime scene samples.

“CODIS,” by the way, is the Combined DNA Index System, a combination of federal, state, and local systems.

“Rapid DNA” is an accelerated, automated DNA method that can process DNA samples in less than two hours, as opposed to the more traditional DNA processes that can take a lot longer.

The FBI is NOT ready to use rapid DNA to solve crimes, although some local police agencies have chosen to do so. And until February of this year, the FBI was not ready to use rapid DNA in the booking process either.

So what has been authorized?

The Bureau recognizes that National DNA Index System (NDIS) approval of the Rapid DNA Booking Systems and training of law enforcement personnel using the approved systems are integral to ensuring that Rapid DNA is used in a manner that maintains the quality and integrity of CODIS and NDIS.

Rapid DNA Booking System(s) approved for use at NDIS by a law enforcement booking station are listed below.

ANDE 6C Series G (effective February 1, 2021)

RapidHIT™ ID DNA Booking System v1.0 (effective July 1, 2021) 

If you read the FBI rapid DNA page, you can find links to a number of forensic, security, and other standards that have to be followed when using rapid DNA in a booking environment.

But those aren’t the only restrictions on rapid DNA use.

Can ANY law enforcement agency use rapid DNA in booking?

Um, no.

According to the National Conference of State Legislatures (2013; see PDF), not all states authorize the taking of DNA after an arrest. As of 2013, 20 states did NOT allow the taking of DNA from individuals who had been arrested but not convicted. And of the 30 remaining states, some (such as Connecticut) only allowed taking of DNA for “serious felonies,” some (such as California) for all felonies, and various mixtures in between. Oklahoma, for example, only allowed taking of DNA for “aliens unlawfully present under federal immigration law.”

Now, of course, a rogue police officer could take your DNA when not legally authorized to do so. Then again, a rogue restaurant employee could put laxatives in your food; that doesn’t mean we outlaw laxatives.

An “investigative lead”

So let’s say that you’re arrested for a crime, and your state allows the taking of DNA for your crime at arrest, and your local law enforcement agency has a rapid DNA instrument.

Now let’s assume that your DNA is searched against a DNA database of unsolved crimes, and your DNA matches a sample from another crime. What happens next?

If there is a match, police will likely want to take a closer look.

Wait a minute. There’s a DNA match! Doesn’t that mean that the police can swoop in and arrest the individual, and the individual is immediately convicted?

Um, no. Stop trusting your TV.

It takes more than DNA to convict a person of a crime.

While DNA can provide an investigative lead, DNA in and of itself is not sufficient to convict an individual. The DNA evidence usually has to be supported by additional evidence.

Especially since there may be other explanations of how the DNA got there.

In 2011, Adam Scott’s DNA matched with a sperm sample taken from a rape victim in Manchester—a city Scott, who lived more than 200 miles away, had never visited. Non-DNA evidence subsequently cleared Scott. The mixup was due to a careless mistake in the lab, in which a plate used to analyze Scott’s DNA from a minor incident was accidentally reused in the rape case.

Then there’s the uncomfortable and inconvenient truth that any of us could have DNA present at a crime scene—even if we were never there. Moreover, DNA recovered at a crime scene could have been deposited there at a time other than when the crime took place. Someone could have visited beforehand or stumbled upon the scene afterward. Alternatively, their DNA could have arrived via a process called secondary transfer, where their DNA was transferred to someone else, who carried it to the scene.

But there is a DNA case that was (originally) puzzling. Actually, a whole bunch of DNA cases.

There is an interesting case, known as the Phantom of Heilbonn, that dates from 1993 in Austria, France and Germany. From that year the DNA of an unknown female was detected at crime scenes in those countries, including at six murder scenes, one of the victims being a female police officer from Heilbronn, Germany. Between 1993 and March 2009 the woman’s DNA was detected at 40 crime scenes which ranged from murder to burglaries and robberies. The DNA was found on items ranging from a biscuit to a heroin syringe to a stolen car.

Then it got really weird.

In March 2009 investigators discovered the same DNA on the burned body of a male asylum-seeker in France. Now this presented something of an anomaly: the corpse was male but the DNA was of a female.

You guessed it; it was the swabs themselves that were contaminated.

So a DNA match is just the start of an investigative process, but it could provide the investigative lead that eventually leads to the conviction of an individual.

Perhaps you’ve noticed that I use the phrase “investigative lead” a lot when talking about DNA and about facial recognition. Trust me, it’s important.

But is the taking of DNA at booking Constitutional?

Obviously this is a huge question, because technical ability to do something does not automatically mean that you are Constitutionally authorized to do so. There is, after all, Fourth Amendment language protecting us against “unreasonable searches and seizures.”

Is the taking of DNA from arrestees who have not been convicted (assuming state law allows it) reasonable, or unreasonable?

Alonzo Jay King, Jr. had a vested interest in this question.

Alonzo Jay King Jr…was arrested in 2009 on assault charges. Before he was convicted of that crime, police took a DNA sample pursuant to Maryland’s new law allowing for such collections at the time of arrest in certain offenses….

I want to pause right here to make sure that the key point is highlighted. King, an arrestee who had not been convicted at the time of any crime, was compelled to provide evidence. At the time of arrest, collection of certain types of evidence (such as fingerprints) is “reasonable.” But collection of certain other types of evidence (such as a forced confession) is “unreasonable.”

So King’s DNA was taken and was searched against a Maryland database of DNA from unsolved crimes. You won’t believe what happened next! (Actually, you will.)

The DNA matched a sample from an unsolved 2003 rape case, and Mr. King was convicted of that crime.

Sentenced to life in prison, actually.

Wicomico County Assistant State’s Attorney Elizabeth L. Ireland said she requested the court impose a life sentence on King, not only because of his past criminal convictions, but also because it turned out that he was a friend of the victim’s family. She said this proved King was a continuing danger to the community.

Before you say, “well, if he was the rapist, he should be imprisoned, legal niceties notwithstanding,” think of the implications of that statement. The entire U.S. legal system is based upon the premise that it is better for a guilty person to mistakenly go free than for an innocent person to mistakenly be punished.

And if that doesn’t sink in…what if YOU were arrested and convicted unlawfully? What if a plate analyzing YOUR DNA wasn’t cleaned properly, and you were unjustly convicted of rape? Or what if a confession were coerced from YOU, and used to convict you?

So King’s question was certainly important, regardless of whether or not he actually committed the rape for which he was convicted.

King therefore appealed on Fourth Amendment grounds, the Maryland Court of Appeals overturned his conviction (PDF), and the State of Maryland brought the case to the U.S. Supreme Court in 2013 (Maryland v. King). In a close 5-4 decision (PDF) in which both conservatives and liberals were on both sides of the argument, the Court ruled that the taking of DNA from arrestees WAS Constitutional.

But that wasn’t the end of the argument, because a new case arose in the state of California. But the California Supreme Court ruled in 2018 that the practice was allowed in that state.

So the taking of DNA at booking is not only authorized (in some states, for some charges), it’s also Constitutional. (Although the Supreme Court’s opinion is still widely debated.)

From http://bytesdaily.blogspot.com/2011/03/buccal-swabs-dna-and-phantoms.html

So anyone who gets arrested for a felony in my home state of California should be ready for a buccal (cheek) swab.

Faulty “journalism” conclusions: the Israeli “master faces” study DIDN’T test ANY commercial biometric algorithms

(Part of the biometric product marketing expert series)

Modern “journalism” often consists of reprinting a press release without subjecting it to critical analysis. Sadly, I see a lot of this in publications, including both biometric and technology publications.

This post looks at the recently announced master faces study results, the datasets used (and the datasets not used), the algorithms used (and the algorithms not used), and the (faulty) conclusions that have been derived from the study.

Oh, and it also informs you of a way to make sure that you don’t make the same mistakes when talking about biometrics.

Vulnerabilities from master faces

From https://www.facebook.com/UplandPolice/photos/pcb.5828261687245780/5828261513912464/

In facial recognition, there is a concept called “master faces” (similar concepts can be found for other biometric modalities). The idea behind master faces is that such data can potentially match against MULTIPLE faces, not just one. This is similar to a master key that can unlock many doors, not just one.

This can conceivably happen because facial recognition algorithms do not match faces to faces, but match derived features from faces to derived features from faces. So if you can create the right “master” feature set, it can potentially match more than one face.

However, this is not just a concept. It’s been done, as Biometric Update informs us in an article entitled ‘Master faces’ make authentication ‘extremely vulnerable’ — researchers.

Ever thought you were being gaslighted by industry claims that facial recognition is trustworthy for authentication and identification? You have been.

The article goes on to discuss an Israeli research project that demonstrated some true “master faces” vulnerabilities. (Emphasis mine.)

One particular approach, which they write was based on Dlib, created nine master faces that unlocked 42 percent to 64 percent of a test dataset. The team also evaluated its work using the FaceNet and SphereFace, which like Dlib, are convolutional neural network-based face descriptors.

They say a single face passed for 20 percent of identities in Labeled Faces in the Wild, an open-source database developed by the University of Massachusetts. That might make many current facial recognition products and strategies obsolete.

Sounds frightening. After all, the study not only used dlib, FaceNet, and SphereFace, but also made reference to a test set from Labeled Faces in the Wild. So it’s obvious why master faces techniques might make many current facial recognition products obsolete.

Right?

Let’s look at the datasets

It’s always more impressive to cite an authority, and citations of the University of Massachusetts’ Labeled Faces in the Wild (LFW) are no exception. After all, this dataset has been used for some time to evaluate facial recognition algorithms.

But what does Labeled Faces in the Wild say about…itself? (I know this is a long excerpt, but it’s important.)

DISCLAIMER:

Labeled Faces in the Wild is a public benchmark for face verification, also known as pair matching. No matter what the performance of an algorithm on LFW, it should not be used to conclude that an algorithm is suitable for any commercial purpose. There are many reasons for this. Here is a non-exhaustive list:

Face verification and other forms of face recognition are very different problems. For example, it is very difficult to extrapolate from performance on verification to performance on 1:N recognition.

Many groups are not well represented in LFW. For example, there are very few children, no babies, very few people over the age of 80, and a relatively small proportion of women. In addition, many ethnicities have very minor representation or none at all.

While theoretically LFW could be used to assess performance for certain subgroups, the database was not designed to have enough data for strong statistical conclusions about subgroups. Simply put, LFW is not large enough to provide evidence that a particular piece of software has been thoroughly tested.

Additional conditions, such as poor lighting, extreme pose, strong occlusions, low resolution, and other important factors do not constitute a major part of LFW. These are important areas of evaluation, especially for algorithms designed to recognize images “in the wild”.

For all of these reasons, we would like to emphasize that LFW was published to help the research community make advances in face verification, not to provide a thorough vetting of commercial algorithms before deployment.

While there are many resources available for assessing face recognition algorithms, such as the Face Recognition Vendor Tests run by the USA National Institute of Standards and Technology (NIST), the understanding of how to best test face recognition algorithms for commercial use is a rapidly evolving area. Some of us are actively involved in developing these new standards, and will continue to make them publicly available when they are ready.

So there are a lot of disclaimers in that text.

  • LFW is a 1:1 test, not a 1:N test. Therefore, while it can test how one face compares to another face, it cannot test how one face compares to a database of faces. The usual law enforcement use case is to compare a single face (for example, one captured from a video camera) against an entire database of known criminals. That’s a computationally different exercise from the act of comparing a crime scene face against a single criminal face, then comparing it against a second criminal face, and so forth.
  • The people in the LFW database are not necessarily representative of the world population, the population of the United States, the population of Massachusetts, or any population at all. So you can’t conclude that a master face that matches against a bunch of LFW faces would match against a bunch of faces from your locality.
  • Captured faces exhibit a variety of quality levels. A face image captured by a camera three feet from you at eye level in good lighting will differ from a face image captured by an overhead camera in poor lighting. LFW doesn’t have a lot of these latter images.

I should mention one more thing about LFW. The researchers allow testers to access the database itself, essentially making LFW an “open book test.” And as any student knows, if a test is open book, it’s much easier to get an A on the test.

By MCPearson – Own work, CC BY-SA 3.0, https://commons.wikimedia.org/w/index.php?curid=25969927

Now let’s take a look at another test that was mentioned by the LFW folks itself: namely, NIST’s Face Recognition Vendor Test.

This is actually a series of tests that has evolved over the years; NIST is now conducting ongoing tests for both 1:1 and 1:N (unlike LFW, which only conducts 1:1 testing). This is important because most of the large-scale facial recognition commercial applications that we think about are 1:N applications (see my example above, in which a facial image captured at a crime scene is compared against an entire database of criminals).

In addition, NIST uses multiple data sets that cover a number of use cases, including mugshots, visa photos, and faces “in the wild” (i.e. not under ideal conditions).

It’s also important to note that NIST’s tests are also intended to benefit research, and do not necessarily indicate that a particular algorithm that performs well for NIST will perform well in a commercial implementation. (If the algorithm is even available in a commercial implementation: some of the algorithms submitted to NIST are research algorithms only that never made it to a production system.) For the difference between testing an algorithm in a NIST test and testing an algorithm in a production system, please see Mike French’s LinkedIn article on the topic. (I’ve cited this article before.)

With those caveats, I will note that NIST’s FRVT tests are NOT open book tests. Vendors and other entities give their algorithms to NIST, NIST tests them, and then NIST tells YOU what the results were.

So perhaps it’s more robust than LFW, but it’s still a research project.

Let’s look at the algorithms

Now that we’ve looked at two test datasets, let’s look at the algorithms themselves and evaluate the claim that results for the three algorithms Dlib, FaceNet, and SphereFace can naturally be extrapolated to ALL facial recognition algorithms.

This isn’t the first time that we’ve seen such an attempt at extrapolation. After all, the MIT Media Lab’s Gender Shades study (which evaluated neither 1:1 nor 1:N use cases, but algorithmic attempts to identify gender and race) itself only used three algorithms. Yet the popular media conclusion from this study was that ALL facial recognition algorithms are racist.

Compare this with NIST’s subsequent study, which evaluated 189 algorithms specially for 1:1 and 1:N use cases. While NIST did find some race/sex differences in algorithms, these were not universal: “Tests showed a wide range in accuracy across developers, with the most accurate algorithms producing many fewer errors.”

In other words, just because an earlier test of three algorithms demonstrated issues in determining race or gender, that doesn’t mean that the current crop of hundreds of algorithms will necessarily demonstrate issues in identifying individuals.

So let’s circle back to the master faces study. How do the results of this study affect “current facial recognition products”?

The answer is “We don’t know.”

Has the master faces experiment been duplicated against the leading commercial algorithms tested by Labeled Faces in the Wild? Apparently not.

Has the master faces experiment been duplicated against the leading commercial algorithms tested by NIST? Well, let’s look at the various ways you can define the “leading” commercial algorithms.

For example, here’s the view of the test set that IDEMIA would want you to see: the 1:N test sorted by the “Visa Border” column (results as of August 6, 2021):

From https://pages.nist.gov/frvt/html/frvt1N.html as of August 6, 2021.

And here’s the view of the test set that Paravision would want you to see: the 1:1 test sorted by the “Mugshot” column (results as of August 6, 2021):

From https://pages.nist.gov/frvt/html/frvt11.html as of August 6, 2021.

Now you can play with the sort order in many different ways, but the question remains: have the Israeli researchers, or anyone else, performed a “master faces” test (preferably a 1:N test) on the IDEMIA, Paravision, Sensetime, NtechLab, Anyvision, or ANY other commercial algorithm?

Maybe a future study WILL conclude that even the leading commercial algorithms are vulnerable to master face attacks. However, until such studies are actually performed, we CANNOT conclude that commercial facial recognition algorithms are vulnerable to master face attacks.

So naturally journalists approach the results critically…not

But I’m sure that people are going to make those conclusions anyway.

From https://xkcd.com/386/. Attribution-NonCommercial 2.5 Generic (CC BY-NC 2.5).

Does anyone even UNDERSTAND these studies? (Or do they choose NOT to understand them?)

How can you avoid the same mistakes when communicating about biometrics?

As you can see, people often write about biometric topics without understanding them fully.

Even biometric companies sometimes have difficulty communicating about biometric topics in a way that laypeople can understand. (Perhaps that’s the reason why people misconstrue these studies and conclude that “all facial recognition is racist” and “any facial recognition system can be spoofed by a master face.”)

Are you about to publish something about biometrics that requires a sanity check? (Hopefully not literally, but you know what I mean.)

Well, why not turn to a biometric content marketing expert? Use the identity/biometric blog expert to write your blog post, the identity/biometric case study expert to write your case study, or the identity/biometric white paper expert to…well, you get the idea. (And all three experts are the same person!)

Bredemarket offers over 25 years of experience in biometrics that can be applied to your marketing and writing projects.

If you don’t have a content marketing project now, you can still subscribe to my Bredemarket Identity Firm Services LinkedIn page or my Bredemarket Identity Firm Services Facebook group to keep up with news about biometrics (or about other authentication factors; biometrics isn’t the only one). Or scroll down to the bottom of this blog post and subscribe to my Bredemarket blog.

If my content creation process can benefit your biometric (or other technology) marketing and writing projects, contact me.

You will soon deal with privacy stakeholders (and they won’t care about the GYRO method)

(Part of the biometric product marketing expert series)

I’ve written about the various stakeholders at government agencies who have an interest in biometrics procurements- not only in this post, but also in a post that is available to Bredemarket Premium subscribers. One of the stakeholders that appeared on my list was this one.

The privacy advocate who needs to ensure that the biometric data complies with state and national privacy laws.

Broken Liberty: Istanbul Archaeology Museum. By © Nevit Dilmen, CC BY-SA 3.0, https://commons.wikimedia.org/w/index.php?curid=1115936

If you haven’t encountered a privacy advocate in your marketing or proposal efforts…you will.

Utah Gov. Spencer Cox has appointed Christopher Bramwell as the Department of Government Operations’ first privacy officer….As privacy officer, Bramwell will be responsible for surveying and compiling information about state agencies’ privacy practices to discern which poses a risk to individual privacy. He will also work with the personal privacy oversight commission and state privacy officer to provide government privacy practice reports and recommendations.

Obviously this affects companies that work with government agencies on projects such as digital identity platforms. After all, mobile driver’s licenses contain a wealth of personally identifiable information (PII), and a privacy advocate will naturally be concerned about who has access to this PII.

But what about law enforcement? Do subjects in law enforcement databases have privacy rights that need to be respected? After all, law enforcement agencies legally share PII all the time.

From https://www.facebook.com/UplandPolice/photos/pcb.5828261687245780/5828261513912464/

However, there are limitations on what law enforcement agencies can share.

  • First off, remember that not everyone in a law enforcement database is an arrested individual. For example, agencies may maintain exclusion databases of police officers and crime victims. When biometric evidence is found at a crime scene, agencies may compare the evidence against the exclusion database to ensure that the evidence does not belong to someone who is NOT a suspect. (This can become an issue in DNA mixtures, by the way.)
  • Second off, even arrested individuals have rights that need to be respected. While arrested individuals lose some privacy rights (for example, prisoners’ cells can be searched and prisoners’ mail can be opened), a privacy advocate should ensure that any system does not deny prisoners protections to which they are entitled.

So expect to see a raised concern about privacy rights when dealing with law enforcement agencies. This concern will vary from jurisdiction to jurisdiction based upon the privacy (and biometric) laws that apply in each jurisdiction, but vendors that do business with government agencies need to stay abreast of privacy issues.

A little more about stakeholders, or actors, or whoever

Whether you’re talking about stakeholders in a government agency, stakeholders at a vendor, or external stakeholders, it’s important to identify all of the relevant stakeholders.

Or whatever you call them. I’ve been using the term “stakeholders” to refer to these people in this post and the prior posts, but there are other common terms that could be used. People who construct use cases refer to “actors.” Marketers will refer to “personas.”

Whatever term you use, it’s important to distinguish between these stakeholders/actors/personas/whatever. They have different motivations and need to be addressed in different ways.

When talking with Bredemarket clients, I often need to distinguish between the various stakeholders, because this can influence my messaging significantly. For example, if a key decision-maker is a privacy officer, and I’m communicating about a fingerprint identification system, I’m not going to waste a lot of time talking about the GYRO method.

My time wouldn’t be wasted effort if I were talking to a forensic examiner, but a privacy advocate just wouldn’t care. They would just sit in silence, internally musing about the chances that a single latent examiner’s “green” determination could somehow expose a private citizen to fraud or doxxing or something.

This is why I work with my clients to make sure that the messaging is appropriate for the stakeholder…and when necessary, the client and I jointly develop multiple messages for multiple stakeholders.

If you need such messaging help, please contact Bredemarket for advice and assistance. I can collaborate with you to ensure that the right messages go to the right stakeholders.

The Surfside building collapse may require a redefinition of “real-time” regarding rapid DNA

(Part of the biometric product marketing expert series)

I’ve previously noted that the definition of “real-time” can vary depending upon the use case. In the automated fingerprint identification systems world of the late 1990s, a definition of “real-time” in minutes was appropriate, but for the computer aided dispatch world, “real-time” was (and is) measured in seconds.

“Hi, SCC folks, welcome to Printrak. You’re joining a company that sells REAL TIME AFIS that delivers results within one minute! Aren’t you impressed?”

“Hello, new corporate overlords. We provide computer aided dispatch systems that send police, fire, and medical personnel to crime scenes and emergency sites as soon as possible. If our CAD systems took AN ENTIRE MINUTE to dispatch personnel, PEOPLE WOULD DIE. We use really powerful computers to get personnel dispatched in a second. Enjoy your real time AFIS…amateurs.”

I also mentioned a two-hour “real-time” use case, which is (conservatively) the time it takes a rapid DNA instrument to do its work.

From the FBI page on Rapid DNA, https://www.fbi.gov/services/laboratory/biometric-analysis/codis/rapid-dna

The rapid DNA vendors provide machines that can perform an automated DNA analysis in 90 minutes, a vast improvement over traditional DNA especially when existing backlogs are taken into account. And for the most part, 90 minutes is fine.

But the Surfside tragedy illustrates how 90 minutes may not be adequate.

There’s already been coverage of how rapid DNA can be, and is being, used to identify victims of the Surfside building collapse. NPR ran an article on this, and WFLA aired a news report.

To date I have not found a public source that lists how many rapid DNA machines are being used in the investigation, but let’s do a little math and see how many rapid DNA instruments could possibly be required.

Assume a conservative two hours is required to fully analyze each DNA sample and determine the possible identity of a deceased victim. Further assume that because of the importance of this case, the DNA instruments are being operated 24 hours a day, 7 days a week. No going home at 5:00 pm in this case, which is receiving international attention.

Now let’s look at the numbers. As of 2:00 pm Eastern Daylight Time today, 20 deaths are confirmed, and 128 people are still unaccounted for.

What happens if there is a sudden horrific discovery of 100 deceased? How long would it take to identify all of them?

If 3 rapid DNA instruments are available, and each is processing 12 DNA samples in a 24 hour day, then it would take about three days to run all the samples through the DNA instruments.

Three very long days for the families of the potential victims who are waiting for news.

So the authorities may need to move to plan B.

The Indian River County Sheriff’s Office has been notified it might be asked to respond with the agency’s rapid DNA test machines to the deadly condominium collapse in Surfside, Sheriff Eric Flowers said….

“They put our folks on standby last weekend to respond if theirs got overwhelmed,” Flowers said. “At this point, they’ve not called for that, but our folks are ready and our machines are ready that if they call us we will respond to assist in DNA identification.”

Yes, in this case you can throw more machines at a problem to solve it, provided that you have the proper personnel to support them. Luckily, the rapid DNA instruments themselves do not need a forensic background to operate them, since they are designed to operate in an automated fashion. However, if rapid DNA analysis has an inconclusive result, then additional traditional DNA analysis will have to be performed which will require forensic expertise. (That, however, is outside of the scope of this post.)

By Zephyris – Own work, CC BY-SA 3.0, https://commons.wikimedia.org/w/index.php?curid=15027555

So where do we stand after Surfside?

Previous rapid DNA identification efforts have just involved one person or less than a dozen people. But this case, in which potentially over 100 people may need to be identified, is truly pushing the limits of the technology.

(Come to think of it, it’s similar to how video analytic analysis was pushed to the limits by the Boston Marathon bombings. But I digress.)

And sadly, there have already been instances in which that many people, or more people, needed to be identified. Imagine, for example, the crash of a large airplane. Or worse still, the crash of two large airplanes into a skyscraper.

And now this 90 minute response time suddenly doesn’t seem so fast any more.

Is your home your castle when you use consumer doorbell facial recognition?

(Part of the biometric product marketing expert series)

For purposes of this post, I will define three entities that can employ facial recognition:

  • Public organizations such as governments.
  • Private organizations such as businesses.
  • Individuals.

Some people are very concerned about facial recognition use by the first two categories of entities.

But what about the third category, individuals?

Can individuals assert a Constitutional right to use facial recognition in their own homes? And what if said individuals live in Peoria?

Concerns about ANY use of facial recognition

Let’s start with an ACLU article from 2018 regarding “Amazon’s Disturbing Plan to Add Face Surveillance to Your Front Door.”

Let me go out on a limb and guess that the ACLU opposes the practice.

The article was prompted by an Amazon 2018 patent application which involved both its Rekognition facial recognition service and its Ring cameras.

One of the figures in Amazon’s patent application, courtesy the ACLU. https://www.aclunc.org/docs/Amazon_Patent.pdf

While the main thrust of the ACLU article concerns acquisition of front door face surveillance (and other biometric) information by the government, it also briefly addresses the entity that is initially performing the face surveillance: namely, the individual.

Likewise, homeowners can also add photos of “suspicious” people into the system and then the doorbell’s facial recognition program will scan anyone passing their home.

I should note in passing that ACLU author Jacob Snow is describing a “deny list,” which flags people who should NOT be granted access such as that pesky solar power salesperson. In most cases, consumer products tout the use of an “allow list,” which flags people who SHOULD be granted access such as family members.

Regardless of whether you’re discussing a deny list or an allow list, the thrust of the ACLU article isn’t that governments shouldn’t use facial recognition. The thrust of the article is that facial recognition shouldn’t be used at all.

The ACLU and other civil rights groups have repeatedly warned that face surveillance poses an unprecedented threat to civil liberties and civil rights that must be stopped before it becomes widespread.

Again, not face surveillance by governments, but face surveillance period. People should not have the, um, “civil liberties” to use the technology.

But how does the tech world approach this?

The reason that I cited that particular ACLU article was that it was subsequently referenced in a CNET article from May 2021. This article bore the title “The best facial recognition security cameras of 2021.”

Let me go out on a limb and guess that CNET supports the practice.

The last part of author Megan Wollerton’s article delves into some of the issues regarding facial recognition use, including those raised by the ACLU. But the bulk of the article talks about really cool tech.

As I stated above, Wollerton notes that the intended use case for home facial recognition security systems involves the creation of an “allow list”:

Some home security cameras have facial recognition, an advanced option that lets you make a database of people who visit your house regularly. Then, when the camera sees a face, it determines whether or not it belongs to someone in your list of known faces. If the recognition system does not know who is at the door, it can alert you to an unknown person on your property.

Obviously you could repurpose such a system for anything you want, provided that you can obtain a clear picture of the face of the pesky social power salesperson.

Before posting her reviews of various security systems, and after a brief mention (expanded later in the article) about possible governmental misuse of facial recognition, Wollerton redirects the conversation.

But let’s step back a bit to the consumer realm. Your home is your castle, and the option of having surveillance cameras with facial recognition software is still compelling for those who want to be on the cutting edge of smart home innovation.

“Your home is your castle” may be a distinctly American concept, but it certainly applies here as organizations such as, um, the ACLU defend a person’s right against unreasonable actions by governments.

Obviously, there are limits to ANY Constitutional right. I cannot exercise my Fourth Amendment right to be secure in my house, couple that with my First Amendment right to freely exercise my religion, and conclude that I have the unrestricted right to perform ritual child sacrifices in my home. (Although I guess if I have a home theater and only my family members are present, I can probably yell “Fire!” all I want.)

So perhaps I could mount an argument that I can use facial recognition at my house any time I want, if the government agrees that this right is “reasonable.”

But it turns out that other people are involved.

You knew I was going to mention Illinois in this post

OK, it’s BIPA time.

As I previously explained in a January 2021 post about the Kami Doorbell Camera, “BIPA” is Illinois’ Biometric Information Privacy Act. This act imposes constraints on a private entity’s use of biometrics. (Governments are excluded in Illinois BIPA.) And here’s how BIPA defines the term “private entity”:

“Private entity” means any individual, partnership, corporation, limited liability company, association, or other group, however organized. A private entity does not include a State or local government agency. A private entity does not include any court of Illinois, a clerk of the court, or a judge or justice thereof.

Did you see the term “individual” in that definition?

So BIPA not only affects company use of biometrics, such as use of biometrics by Google or by a theme park or by a fitness center. It also affects an individual such as Harry or Harriet Homeowner’s use of biometrics.

As republished at https://treeandneighborlawblog.com/2020/02/21/case-of-the-day-friday-february-21-2020/. Fair use.

As I previously noted, Google does not sell its Nest Cam “familiar face alert” feature in Illinois. But I guess it’s possible (via location spoofing if necessary) for someone to buy Nest Cam familiar face alerts in Indiana, and then sneak the feature across the border and implement it in the Land of Lincoln. But while this may (or may not) get Google off the hook, the individual is in a heap of trouble (should a trial lawyer decide to sue the individual).

Let’s face it. The average user of Nest Cam’s familiar face alerts, or the Kami Doorbell Camera, or any other home security camera with facial recognition (note that Amazon currently is not using facial recognition in its consumer products), is probably NOT complying with BIPA.

A private entity in possession of biometric identifiers or biometric information must develop a written policy, made available to the public, establishing a retention schedule and guidelines for permanently destroying biometric identifiers and biometric information when the initial purpose for collecting or obtaining such identifiers or information has been satisfied or within 3 years of the individual’s last interaction with the private entity, whichever occurs first.

I mean it’s hard enough for Harry and Harriet to get their teenage son to acknowledge receipt of the Homeowner family’s written policy for the use of the family doorbell camera. And you can forget about getting the pesky solar power salesperson to acknowledge receipt.

So from a legal perspective, it appears that any individual homeowner who installs a facial recognition security system can be hauled into civil court under BIPA.

But will these court cases be filed from a practical perspective?

Probably not.

When a social media company violates BIPA, the violation conceivably affects millions of individuals and can result in millions or billions of dollars in civil damages.

When the pesky solar power salesperson discovers that Harry and Harriet Homeowner, the damages would be limited to $1,000 or $5,000 plus relevant legal fees.

It’s not worth pursuing, any more than it’s worth pursuing the Illinois driver who is speeding down the expressway at 66 miles per hour.

What is an “antimicrobial” contact fingerprint reader? And what is it NOT?

(Part of the biometric product marketing expert series)

In the COVID and (soon) post-COVID area, people don’t want to touch things. That impacts how identity products are marketed, including biometric readers.

Why contactless biometrics are “better” than contact biometrics

In the biometric world, this reluctance to touch things has served to promote CONTACTLESS biometric technologies, such as facial recognition, other other technologies. The loser in this has been fingerprint-based technologies, as several facial and iris vendors have made the claim that face/iris biometrics are contactless, while fingerprint biometrics are NOT contactless.

Well, my friends at my former employer IDEMIA might take issue with that claim, since you literally do NOT touch the fingerprint reader in IDEMIA’s MorphoWave product. IDEMIA does not (to my knowledge) make any medical claims about MorphoWave, but the company does emphasize that its contactless fingerprint reader allows for fast capture of four-finger slaps.

To protect their premises, organizations need access control solutions that are efficient, fast, and convenient. A contactless fingerprint scanner provides an optimum answer high throughput workplaces. IDEMIA’s MorphoWave™ contactless fingerprint solution scans and verifies 4 fingerprints in less than 1 second, through a fully touchless hand wave gesture. Thanks to the simplicity of this gesture, the throughput can reach up to 50 people per minute.

An antimicrobial contact fingerprint reader?

But what if there were a CONTACT solution that allowed you to capture prints with a reduced fear of “bad things”?

That’s what Integrated Biometrics appears to be claiming.

Integrated Biometrics (IB), the world leader in mobile, FBI-certified biometric fingerprint scanners, and NBD Nanotechnologies (NBD Nano), the surface coating experts, today announced the inclusion of NBD’s RepelFlex MBED transparent coating on IB’s entire line of fingerprint scanners.

An ultra-thin, transparent coating, RepelFlex MBED is designed to provide outstanding antimicrobial, anti-scratch, and anti-stain protection to devices. Long-lasting and multi-functional, RepelFlex MBED is ideal for surfaces that must stand up to high throughput and harsh conditions without compromising accuracy.

So what exactly does “antimicrobial” mean?

A cluster of Escherichia coli bacteria magnified 10,000 times. By Photo by Eric Erbe, digital colorization by Christopher Pooley, both of USDA, ARS, EMU. – This image was released by the Agricultural Research Service, the research agency of the United States Department of Agriculture, with the ID K11077-1 (next)., Public Domain, https://commons.wikimedia.org/w/index.php?curid=958857

Let’s see how NBD Nano describes it.

Preventing the presence and growth of microbials on surfaces is becoming increasingly important. Antimicrobial performance is especially critical on surfaces that are accessible to the public in order to prevent the spread of stain and odor causing bacteria and microbes.

And if you drill further down in NBD Nano’s website, you find this information in a technical data sheet (PDF).

Antimicrobial Performance: Japanese Industrial Standard (JIS) Z 2801 – PASS*
*as tested by Microchem Laboratory, Round Rock, TX

Now since I’m not up to date on my Japanese Industrial Standards, I had to rely on the good folks at the aforementioned Microchem Laboratory to explain what the standard actually means.

The JIS Z 2801 method tests the ability of plastics, metals, ceramics and other antimicrobial surfaces to inhibit the growth of microorganisms or kill them. The procedure is very sensitive to antimicrobial activity and has a number of real world applications anywhere from the hospital/clinical environment to a household consumer company concerned with the ability of a material they have to allow bacterial growth.

The JIS Z 2801 method is the most commonly chosen test and has become the industry standard for antimicrobial hard surface performance in the United States.

It may be antimicrobial, but what about preventing the “C” word?

Now you may have noticed that Microchem Laboratory, NBD Nano, and Integrated Biometrics did not make any medical claims regarding their products. None of them, for example, used the “C” word in any of their materials.

There’s a very, very good reason for that.

If any of these product providers were to make specific MEDICAL claims, then any sales in the United States would come under the purview of the U.S. Food and Drug Administration.

This is something that temperature scanner manufacturers learned the hard way.

Digression: if fever scanners are fever scanners, does that mean they are fever scanners?

Remember “fever scanners”? Those devices that were (and in some cases still are) pointed at your forehead as you enter a building or another secure area? I won’t get into the issues with these devices (what happens when the scanner is placed next to a building’s front entrance on a hot day?), but I will look at some of the claims about those scanners.

About a year ago, John Honovich of IPVM began asking some uncomfortable questions about the marketing of those devices, especially after the FDA clarified what thermal imaging systems could and could not do.

When used correctly, thermal imaging systems generally have been shown to accurately measure someone’s surface skin temperature without being physically close to the person being evaluated….

Thermal imaging systems have not been shown to be accurate when used to take the temperature of multiple people at the same time. The accuracy of these systems depends on careful set-up and operation, as well as proper preparation of the person being evaluated….

Room temperature should be 68-76 °F (20-24 °C) and relative humidity 10-50 percent….

The person handling the system should make sure the person being evaluated…(h)as waited at least 15 minutes in the measurement room or 30 minutes after exercising, strenuous physical activity, bathing, or using hot or cold compresses on the face.

Let’s stop right there. For any of you who have undergone a temperature scan in the last year: how many of you have waited in a measurement room for at least 15 minutes BEFORE your temperature was taken?

Last summer I had a dentist appointment. My dentist is in Ontario, California, where the summers can get kind of hot. The protocol at this dentist’s office was to have you call the office from your car when you arrived in the parking lot, then wait for someone from the office to come outside and take your temperature before you could enter the building.

I was no dummy. I left my car and its air conditioner running while waiting for my temperature to be taken. Otherwise, who knows what my temperature reading would have been? (I also chose NOT to walk to the dentist’s office that day for the same reason.)

Back to John Honovich. He had read the FDA advice on the medical nature of thermal imaging systems, and then noted that some of the manufacturers of said systems were sort of getting around this by stating that their devices were not medical devices.

Even though the manufacturers still referred to them as “fever cameras.”

For example, one vendor (who has since changed its advertising) declared at the time that “thermal temperature-monitoring technology assists in reducing the spread of viral diseases,” even though that vendor’s device “is not a medical device and is not designed or intended for diagnosis, prevention, or treatment of any disease or condition.”

Fever scanners, testosterone supplements…and fingerprint readers

Yes, that language is similar to the language used by providers of natural supplements that, according to anecdotal evidence, work wonders. The FDA really polices this stuff.

So you really don’t want to make medical claims about ANY product unless you can back them up with the FDA. You can say that a particular product passed a particular antimicrobial standard…but you’d better not say anything else.

In fact, Integrated Biometrics only mentions the “antimicrobial” claim in passing, but spends some time discussing other benefits of the NBD Nano technology:

The inclusion of RepelFlex MBED coatings enable IB’s scanners to deliver an even higher level of performance. Surfaces are tougher and more difficult to scratch or stain, increasing their longevity while maintaining print quality even when regular cleaning is not possible due to conditions or times of heavy use.

So the treated Integrated Biometrics products are tough…like those famous 1970s crime fighters Kojak, Columbo, and Danno and the other people from Five-O. (Not that Sherlock and Watson were slouches.)

Book ’em, Danno! By CBS Television – eBay item photo front photo back, Public Domain, https://commons.wikimedia.org/w/index.php?curid=19674714

When biometric readers are “magic” (it’s a small face after all)

(Part of the biometric product marketing expert series)

The news coming across the wire is that Disney’s Magic Kingdom in Florida is testing facial recognition. (H/T International Biometrics + Identity Association.)

“At Walt Disney World Resort, we’re always looking for innovative and convenient ways to improve our guests’ experience—especially as we navigate the impact of COVID-19. With the future in mind and the shift in focus to more touchless experiences, we’re conducting a limited 30-day test using facial recognition technology.”

If the test is successful and facial recognition is implemented, it would be a replacement for (touch) fingerprint technology, which the Disney parks suspended last July for health reasons. (Although touchless fingerprint options are available.)

Disney’s biometric history extends back to 2006, when it used hand geometry.