entropyseal.com/

Neat tamper-evidence packaging concept: a random pattern of beads that's modified when you open the package, along with a smartphone app to automatically verify the pattern is correct.

Reusable glitter basically: to fake a seal, you need to make the significant effort of positioning each bead in the right place. Possible. But very time consuming.

I'd buy one!

Follow

@pete

@giszmo @kekcoin @jgettbtc @121
@jon

today we published a video about the functionality of the current entropyseal prototype:

vimeo.com/508016683
entropyseal.com

@fabianwyss I miss the part in the video where your app tells you if the jar was opened or not.

* close
* seal with app
* check seal with app: succeeds for a range of angles and lighting
* break seal
* check seal with app: fails with display of pattern

Not sure how far in the product cycle you are but I would find it extra cool if the app would recognize the orientation of the lid and normalize it and when checking, show a radar scanner animation superposing old and new scan.

@fabianwyss I imagine the app having a "seal" button which puts it into scan mode. As soon as it recognizes the lid in a good resolution, sharpness and lighting, it freezes the video stream, cuts the image to an ellipse and rotates that ellipse to a default orientation circle with max contrast (lid needs a marker for that).

The "check" button does the same but then diffs both images radar style for one round, displaying the pixel diff in red, with the result expressed as %% match.

@giszmo

Thanks for your valuable inputs.

We have not started with the App yet, focus until here was solely on the hardware.

It is our plan to normalize the image with the help of markers, rotation and homography. Instead of an algorithm to check on similarity, one concept we have in mind is to make a "blinking comparator" that shows normalized reference image and image to verify after each other with adjustable strobe speed. A speed slider helps to check on both images are authentic. If there is a difference, the user can see it.

@fabianwyss So the idea is to let the user judge. For a start that's probably good enough.

@giszmo

The plan is to offer the user a simple and utmost failproof App to visually compare. Algorithm based (dis)similarity in %, highlighting differences in red dots/areas, particle count etc. can be added on top. From a security perspective we believe it is safer to have a simple visual base layer in the App than fancy value prints.

I would go so far one day to show the user 5 sec movieclips of the taken image being automatically rotated to north, un-distorted, scaled, cut in a square standard and bw thresholded. Manipulated algorithm based pattern change would be visually detectable in a movieclip. Wysiwyg motion instead of an obscure algo in the background that is difficult to know what it exactly does and how it can be manipulated backdoor by an attacker.

Sign in to participate in the conversation
Qoto Mastodon

QOTO: Question Others to Teach Ourselves
An inclusive, Academic Freedom, instance
All cultures welcome.
Hate speech and harassment strictly forbidden.