What we collect!

 

Stamporama Discussion Board Logo
For People Who Love To Talk About Stamps
Discussion - Member to Member Sales - Research Center
Stamporama Discussion Board Logo
For People Who Love To Talk About Stamps
Discussion - Member to Member Sales - Research Center
Stamporama Discussion Board Logo
For People Who Love To Talk About Stamps



What we collect!
What we collect!


General Philatelic/Gen. Discussion : Sci News article on detection of counterfeit stamps

 

Author
Postings
cocollectibles

05 Sep 2013
08:50:38am
Here's an interesting article from Science News about detecting counterfeit (forged, etc.) stamps.

Sci News

Cheers,
Peter
Like
Login to Like
this post

"TO ERR IS HUMAN; TO FORGIVE, CANINE."
smaier
Members Picture


Sally

05 Sep 2013
03:18:09pm
re: Sci News article on detection of counterfeit stamps

That's very interesting. However I don't know too many people who have an infrared spectrometer at their disposal. Maybe someday there will be handheld models available for ordinary folks....kind of like digital microscopes today.

Thanks for sharing this.

Sally

Like
Login to Like
this post
Rhinelander
Members Picture


Support the Hobby -- Join the American Philatelic Society

05 Sep 2013
03:23:55pm
re: Sci News article on detection of counterfeit stamps

Thanks Peter for providing this link. It is interesting stuff. We all hope that some technological advancement, a pocket spectrometer or whatever, will take all of the guess work out identifying forgeries. The link only gives the abstract, but I obtained the full text article. My verdict: unimpressive.

The authors perform spectrometric analyses on a sample of Italian stamps from 1862 to 2009 and record the light wave responses.

They first "find" that the components used in the stamp production fall along specific timelines, for instance certain synthetic poly-vinyl acetate adhesive gum does not appear before 1974, or kaolin as a filler material in paper does not appear until 1896 etc. I lifted this image from the paper which displays the timeline of the components used in Italian stamps the researchers "discovered:"

Image Not Found

These eight components, if the stamp used animal adhesive, or not, kaolin coating, or not, are coded in a zero/one fashion, and run through a statistical method called principal components analysis. The method is employed here as a classification tool. The next "finding:" stamps from the same time period tend to group in the same corner of the multi-dimensional statistical space.

Next comes the application: two stamps in the sample did not group with their "peers" and, thus, are subsequently identified as fake or altered.

The first is a re-gumming case. An 1865 stamp had a PVCA gum which did not appear until 1974. The scholars note "This discrepancy is impossible to catch even by the eye of an expert estimator." – I am pretty sure that knowledgeable collectors are quite able to tell re-gumming. But more important is the flip-side of the argument made by the scientists: The only reason that the re-gumming was caught is because the specific adhesive did not exist at the time. In other words, the proposed method does not help if in-period materials are used (i.e., a contemporaneous forgery or alteration).


The second is a forgery of the famous Gronchi Rosa. The stamp is identified as fake due to a different composition compared to the original. The scientist in this case only argue that "their approach constitutes a further scientific tool to confirm that the stamp was forged." In fact, the fake stamp was already certified as a forgery by a philatelic expert "belonging to a well-known set of faked stamps."

So, I am not sure how much the high-powered analyses really contributed in the case of the Gronchi Rosa. Speaking from a strictly scientific/statistical position: The article reproduces the spectrometer readings for the original and fake stamp side-by-side. The "test" that the stamps are not the same consists in an "eye balling" that the pattern of the spectrometer reading is visibly different. Because of a difference in kaolin concentration there are different peaks and valleys.

However, I would not make too much of a sample of "one each." Perhaps one would have to analyze a bunch of originals and fakes, because not every stamp likely will have the exact same reading. If the reality is a certain typical "bandwidth of kaolin" in originals versus forgeries, the next step would be to show that these "bands" will never overlap with a certain statistical probability. That is taking into account soaking in water, benzine, exposure to sun light etc. etc.

In any event, a good read, but, unfortunately, I don't see a panacea for forgery identification here.

Arno

Like
Login to Like
this post
        

 

Author/Postings
cocollectibles

05 Sep 2013
08:50:38am

Here's an interesting article from Science News about detecting counterfeit (forged, etc.) stamps.

Sci News

Cheers,
Peter

Like
Login to Like
this post

"TO ERR IS HUMAN; TO FORGIVE, CANINE."
Members Picture
smaier

Sally
05 Sep 2013
03:18:09pm

re: Sci News article on detection of counterfeit stamps

That's very interesting. However I don't know too many people who have an infrared spectrometer at their disposal. Maybe someday there will be handheld models available for ordinary folks....kind of like digital microscopes today.

Thanks for sharing this.

Sally

Like
Login to Like
this post
Members Picture
Rhinelander

Support the Hobby -- Join the American Philatelic Society
05 Sep 2013
03:23:55pm

re: Sci News article on detection of counterfeit stamps

Thanks Peter for providing this link. It is interesting stuff. We all hope that some technological advancement, a pocket spectrometer or whatever, will take all of the guess work out identifying forgeries. The link only gives the abstract, but I obtained the full text article. My verdict: unimpressive.

The authors perform spectrometric analyses on a sample of Italian stamps from 1862 to 2009 and record the light wave responses.

They first "find" that the components used in the stamp production fall along specific timelines, for instance certain synthetic poly-vinyl acetate adhesive gum does not appear before 1974, or kaolin as a filler material in paper does not appear until 1896 etc. I lifted this image from the paper which displays the timeline of the components used in Italian stamps the researchers "discovered:"

Image Not Found

These eight components, if the stamp used animal adhesive, or not, kaolin coating, or not, are coded in a zero/one fashion, and run through a statistical method called principal components analysis. The method is employed here as a classification tool. The next "finding:" stamps from the same time period tend to group in the same corner of the multi-dimensional statistical space.

Next comes the application: two stamps in the sample did not group with their "peers" and, thus, are subsequently identified as fake or altered.

The first is a re-gumming case. An 1865 stamp had a PVCA gum which did not appear until 1974. The scholars note "This discrepancy is impossible to catch even by the eye of an expert estimator." – I am pretty sure that knowledgeable collectors are quite able to tell re-gumming. But more important is the flip-side of the argument made by the scientists: The only reason that the re-gumming was caught is because the specific adhesive did not exist at the time. In other words, the proposed method does not help if in-period materials are used (i.e., a contemporaneous forgery or alteration).


The second is a forgery of the famous Gronchi Rosa. The stamp is identified as fake due to a different composition compared to the original. The scientist in this case only argue that "their approach constitutes a further scientific tool to confirm that the stamp was forged." In fact, the fake stamp was already certified as a forgery by a philatelic expert "belonging to a well-known set of faked stamps."

So, I am not sure how much the high-powered analyses really contributed in the case of the Gronchi Rosa. Speaking from a strictly scientific/statistical position: The article reproduces the spectrometer readings for the original and fake stamp side-by-side. The "test" that the stamps are not the same consists in an "eye balling" that the pattern of the spectrometer reading is visibly different. Because of a difference in kaolin concentration there are different peaks and valleys.

However, I would not make too much of a sample of "one each." Perhaps one would have to analyze a bunch of originals and fakes, because not every stamp likely will have the exact same reading. If the reality is a certain typical "bandwidth of kaolin" in originals versus forgeries, the next step would be to show that these "bands" will never overlap with a certain statistical probability. That is taking into account soaking in water, benzine, exposure to sun light etc. etc.

In any event, a good read, but, unfortunately, I don't see a panacea for forgery identification here.

Arno

Like
Login to Like
this post
        

Contact Webmaster | Visitors Online | Unsubscribe Emails | Facebook


User Agreement

Copyright © 2024 Stamporama.com