Each code carries individually configured bit-block manipulations. The scan evaluates that structure – not just a serial number.
Are holograms more secure than printed codes?
Printed does not mean insecure. Holographic does not mean uncopyable. Whoever puts all printed security features in one basket is comparing manufacturing processes, not security processes.
"A good printer beats a phone camera."
Some arguments against printed security reduce authentication to a race between printers and smartphone cameras: if a printer can reproduce the image, the copy passes.
That skips the decisive question: what must a counterfeit reproduce to pass authentication? A visible effect? An individual security structure? A valid identity linked to one specific product?
That is where the comparison should begin.
Printed at the limit.
Copied at the limit.
A copy detection pattern is a fine noise field printed into or next to a code at the printer's maximum resolution. The scan checks whether detail was lost – the assumption being that a copy always blurs. Researchers at the University of Geneva tested that assumption with machine learning:
A successful attack exposes a verification method. It does not discredit every printed security technology.
Research has demonstrated machine-learning attacks against specific copy detection patterns and detection methods – and better detection methods in response.2 The relevant question is always which features were reconstructed, which checks were defeated, and under what conditions. Applying those results to every printed security feature is a leap the evidence does not support.
The holographic effect
is only the beginning.
An embossed hologram is a manufactured optical structure – made from a metal master, and reproducible with the same techniques: lift the relief, plate it, emboss it again.3 Its appearance alone does not establish that a label is original.
Wherever a hologram is checked by eye, appearance is all a counterfeit needs: in one tested batch of seized excise holograms, 870 of 894 were fake.4
What about app-verified holographic labels? A vendor of such labels writes in its own patent about the base features of optical security:
"Such security features may, in the case of complete knowledge, be replicated at least up to such an extent that the replicate stands up to an examination by means of a – comparatively coarsely resolving – camera."US 11,390,107 B2, Summary of the invention5
The patent distinguishes these reproducible base features from randomly selected portions, used to make individual labels harder to replicate – in its words, to "aggravate replication". That distinction matters: the security argument depends on a specific construction and its verification – not on the word "hologram". Which reproductions the complete process actually rejects is the question every buyer should ask.
How well automated hologram checks hold up is an open research question. A 2026 study on hologram verification notes that existing methods' "capacity to detect real, dynamic fraud cases (e.g. handcrafted hologram) has not been evaluated to date" and that few of them "can generalize to new, unseen attacks".6
What it means for a brand: a foil label on every unit, applied separately from your artwork, an app for whoever checks – and a security claim that rests on how that specific system was tested.
Reproduce the appearance.
Then test the authenticity.
Crypto-Optical Technology combines individually configured bit-block structures, cryptographic identity checks and scan monitoring. Reproducing a visible code is only one part of the challenge: the security structure must pass verification, the identity must match its assigned product, and suspicious reuse must be detected.




The industrial reprint in our Sample Kit reproduces the appearance – but fails verification: the scan says Not original. Scan the original and the reprint yourself.
ML clone – standard metrics could not reliably tell it apart.
Appearance reproducible – authenticity depends on the specific system.
Industrial reprint – fails verification. Test it with the Sample Kit.
A printer's resolution
answers none of these questions.
Crypto-Optical Technology does not rest on how sharp a print is, or how a surface glitters. Three layers of protection:

The code's identity is tied to its product data, with additional cryptographic checks in the verification.
Every scan is recorded with time and place. Unusual repeated use of the same identity is flagged for investigation.
The glitter film on our sticker is an additional visual feature – the verification does not depend on it. Any phone camera, no app. Printed in standard CMYK by authentic.network through a qualified industrial printing partner – into your label or as a ready-to-apply sticker.
Six questions to ask
any vendor – including us.
- Which reproductions and attack methods were tested?A photocopy is not an attack. An industrial reprint, a machine-learning clone and a plated copy are.
- Who ran the test – the vendor, or someone independent?
- Is an individual security feature verified, or only a valid identifier?
- How often were counterfeits accepted – and originals rejected?
- Does everyone who checks need the vendor's app?Customers, retailers, workshops, customs.
- What does every unit need extra – and how are reuse and product mismatches handled?