Tag Archives: facial recognition
A Look at Biometric Accuracy in an Alternative Universe
(Part of the biometric product marketing expert series)
Imagine an alternative universe in which a single human body had ten (different) faces and only one finger.
- How accurate would facial recognition be?
- How accurate would fingerprint identification be?
Think about the ramifications.
Credit for this thought, not original to me, must remain anonymous.
The Facial Recognition Vendor Is Not At Fault If You Don’t Upgrade Your Software, December 2025 Edition
This is the third time that I’ve seen something like this, so I thought I’d bring attention to it.
Biometric Update recently published a story about a United Kingdom agency that was criticized for its use of Cognitec facial recognition software.
Why? Because the facial recognition software the agency has is not accurate enough, particularly in regards to demographic bias.
Note “the facial recognition software the agency has.” There’s a story here.
“Cognitec released its FaceVACS-DBScan 5.5 software for biometrics matching at scale in 2020….The current version is 5.9, but Home Office’s Police National Database uses 5.5, which is why that version was tested.”
Important clarification.
Now perhaps the agency had its reasons for not upgrading the Cognitec software.
But governments and enterprises should not use old facial recognition software. Unless they have to run the software on computers running PC-DOS. Then they have other problems.
And if you detected that this post sounds really really similar to one I wrote back in April…you’re right. Back then an Australian agency continued to use an older version of the Cognitec algorithm, even though a newer one was available.
But I’m still using the pre-Nano Banana illustration for this new post.
A question for you: is YOUR company using outdated content? Are you ready to update it? Talk to Bredemarket.
The Biometric Challenges of Twins
Facial recognition and DNA can’t tell identical twins apart. (Well, not unless you can detect somatic mutations.) Even Santa can’t tell.
And fingerprints don’t work because everyone at the North Pole wears gloves.
Face Product Marketing Expert (27 posts)
To ensure that my social media followers don’t have all the fun with my “biometric product marketing expert” shares, here are links to some Bredemarket blog posts on facial recognition (identification) and facial analysis (classification).
Facial recognition:
- (10/14/2025) Identity and Expression.
- (10/7/2025) Amazon’s Take on “Familiar Faces” is Not Available Everywhere.
- (9/27/2025) Active Liveness Detection.
- (8/29/2025) Forget About Milwaukee’s Facial Recognition DATA: We All Want to See Milwaukee’s Facial Recognition POLICY.
- (5/13/2025) Who Tests Facial Recognition Algorithms?
- (5/9/2025) Revisiting Amazon Rekognition, May 2025.
- (4/25/2025) Is Milwaukee Selling PII for Free Facial Recognition Software Access?
- (4/18/2025) The Facial Recognition Vendor Is Not At Fault If You Don’t Upgrade Your Software.
- (3/14/2025) Know Your…Passenger.
- (2/13/2025) How to Recognize People From Quite a Long Way Away.
- (12/19/2024) Hospital Patient Facial Recognition.
- (5/31/2024) The Why, How, and What on NIST Age Estimation Testing.
- (5/23/2024) Positioning, Messaging, and Your Facial Recognition Product Marketing.
- (1/23/2024) If We Don’t Train Facial Recognition Users, There Will Be No Facial Recognition.
- (9/26/2023) More on NIST’s FRTE-FATE Split.
- (6/5/2023, from jebredcal) The Truth About NIST Facial Recognition Sub-Tests: Which Ones Are Most Important for YOUR Business?
- (2/23/2022) Who is THE #1 NIST facial recognition vendor?
- (8/6/2021) Faulty “journalism” conclusions: the Israeli “master faces” study DIDN’T test ANY commercial biometric algorithms.
- (7/1/2021) Is your home your castle when you use consumer doorbell facial recognition?
- (6/22/2021, from LinkedIn) Don’t ban facial recognition.
- (3/24/2021) When biometric readers are “magic” (it’s a small face after all).
- (10/21/2020, from jebredcal) Who can, and who can’t, use facial recognition in Portland, Oregon.
Facial analysis:
- (6/3/2025) The Monk Skin Tone Scale.
- (3/28/2025) Age Estimation is Challenging.
- (9/22/2023, from LinkedIn) Why NIST Cares About Presentation Attack Detection…and Why You Should Also.
- (7/3/2023) I Changed My Mind on Age Estimation.
- (9/7/2020, from jebredcal) Classification, Individualization, and Albino Wallabies.
Bad To The Bone
Your residential biometric readers (outside of Illinois) may encounter challenges on Friday, October 31.
Face and Voice, the Popular Combination
Face and voice, the popular combination. At least on LinkedIn. Maybe you can fool one biometric modality, but it’s much harder to fool two.
(Picture from Google Gemini)
Sometimes You Don’t Need, Or Want, Identity Assurance Level 3 (IAL3)
This post is specifically for firms that sell identity verification solutions at various identity assurance levels, or IALs.
I have written a post entitled “Identity Assurance Level 3 (IAL3): When Identity Assurance Level 2 (IAL2) Isn’t Good Enough.”
Which naturally implies that IAL3 is better than IAL2, because it’s more secure.
So why doesn’t EVERYONE use IAL3?
For the same reason that childrens’ piggy banks aren’t protected with multiple biometric modalities AND driver’s license authentication.
Kids don’t have driver’s licenses anyway.
In the same vein, in-person or remote supervised identity proofing isn’t always necessary. If your business would lose customers by insisting upon IAL3, and you’re OK with assuming the financial risk, don’t do it.
Imagine if you had to get on a video chat and show your face and your driver’s license before EVERY Amazon purchase. Customers would go elsewhere. Amazon would go broke within days.
Which is why some identity firms promote IAL3, while others promote IAL2. (I won’t talk about the firms that promote IAL1.)

Whatever identity assurance level your prospects need, Bredemarket can help you create the content. Let’s talk about your specific needs.
Caught!
I was having fun creating videos based upon the controversial third verse of The Star Spangled Banner, but I decided to get back to business.
And the business is that, as the Innocence Project knows all too well, algorithms can be better than humans at identifying faces.
But the silly videos are only what I do for fun.
What I do for business is help identity, biometrics, and technology companies explain how their solutions benefit society.
Can Bredemarket help YOUR firm come up with the right words, via compelling content creation?
- Blog posts. Among other projects, I’ve authored a multi-month blog series to attract business to a client.
- Case studies and testimonials. Among other projects, I’ve written a dozen case studies to justify a firm’s capabilities to its projects.
- LinkedIn articles and posts. The multi-month blog series was designed for repurposing as LinkedIn articles.
- White papers. My white papers have made the case for the superiority of my clients’ products and services.
Set up a free meeting to talk to Bredemarket about your marketing and writing needs.
Identity and Expression
(Part of the biometric product marketing expert series)
Whether you are a human or a non-person entity (NPE) with facial recognition capability, you rely on visual cues to positively identify or authenticate a person. Let’s face it; many people resemble each other, but specific facial expressions or emotions are not always shared by people who otherwise look alike.

But in one of those oddities that fill the biometric world, you can have TOO MUCH expression. Part 3 of International Civil Aviation Organization (ICAO) Document 9303, which governs machine readable travel documents, mandates that faces on travel documents must maintain a neutral expression without smiling. At the time (2003) it was believed that the facial recognition algorithms would work best if the subject were expressionless. I don’t know if that holds true today.

But once the smile is erased, any other removal of expression or emotion degrades identification capability significantly. For example, closing the eyes not only degrades facial recognition, but is obviously fatal to iris recognition.

And if you remove the landmarks upon which facial recognition depends, identification is impossible.

While expression or lack thereof does not invalidate the assumption of permanence of the biometric authentication factor, it does govern the ability of people and machines to perform identification or authentication.
