Blob Detection Counterfeit Technology
Counterfeiting higher-value circulating coins can be not only profitable but challenging to detect. AI Technology could be the answer to the problem.
Make no mistake, counterfeiters are still faking coins as well as bank notes. Counterfeiting higher-value circulating coins can be not only profitable but challenging to detect since the average person is unlikely to take a close look at their “change,” while the feel of a bank note might be sufficient to draw suspicion to what may prove to be a fake.
Counterfeit coin discoveries don’t always get the attention they deserve. Numismatic Guaranty Corporation has reported 2023 Silver American Eagles composed of 65 percent copper, 21 percent zinc and 13 percent nickel, with less than a half percent of silver added to this mix. Fortunately, these fakes aren’t too challenging to detect. The planchet is too thick, the security notch is in the wrong place, the coin is about a dozen notches short on the edge reeding, the luster is off, there are design problems, and the weight is incorrect. Nevertheless, at least one example got as far as this third-party authentication service before being pronounced to be bogus.
In 2023, Jean-Francois Généreux, from Sorel, Quebec in Canada, was reported to have been arrested following when a Canada Border Services Agency customs officer seized 26,000 alleged fake 2-dollar coins at a FedEx warehouse at the Montreal-Mirabel International Airport. Généreux had another 14,000 fakes in his home. All were dated 2012, and all originated from China. The queen’s nose is too long and sharp, a maple leaf-shaped security device that doesn’t exist on genuine 2012 coins was present, and other design problems were identified on the illicit imports.
The Pontifical Catholic University of Peru in Lima recently purchased a batch of historical coins, including an 1899 Peruvian dinero, from a local coin dealer for university chemistry graduate students to study. The only problem is that Peru didn’t make any dinero coins in 1899. In a paper published about the study of that specific coin, Dr. Luis Ortega-San-Martin commented, “The refining methods were not as good as they are now.” It has been suggested the coin was likely produced during the late 19th and early 20th century, adding, “There must be a few [more] around.” Since others have not been reported, it is possible such fakes may be in coin collections, according to Ortega-San-Martin.
There have been so many counterfeit 2-euro coins circulating in Kosovo that an estimated 2.5 million coins of this denomination have been withdrawn by the nation’s central bank. On May 14 Kosovar police authorities estimated about 54,000 fakes had been detected. EuroNews.com reported, “The phenomenon has caused many businesses and traders to refuse payments with the currency due to the mass spread of counterfeit currencies.”
In April, Spain’s national police broke up a ring of counterfeiters distributing about 500,000 “high quality” fake 2-euro coins into circulation all across Europe. Following up on this, Germany’s Bundesbank or central bank posted an online page titled Leitfaden Münzen (Guidelines Coins) that states, “Devices that have been successfully tested are available on the website of the European Commission.”
On April 27, DW (Deutsche Welle) commented, “Of course, this doesn’t help the common person much. So, the bank recommends paying attention to ‘the first impression.’ On a real coin, the imprinted image clearly rises above the rest of the coin’s surface, and all contours can be clearly seen, the bank writes. Caution is called for if this isn’t the case.”
Researchers at the Centre for Pattern Recognition and Machine Intelligence AT Concordia University in Montreal, Canada, recently published a paper in the journal Expert Systems With Applications that presents a novel framework using image-mining techniques and machine learning algorithms to identify flaws in counterfeit coins.
Department of Computer Science and Software Engineering Professor Ching Suen said, “Using image technology, we scanned both genuine and counterfeit coins so we can look for anomalies that may be either two- or three-dimensional, such as letters or the face of the person on the coin.”
CENPARMI Postdoctoral Fellow and lead author of the paper Maryam Sharifi Rad added, “This framework is not only about safeguarding our economy and resources… It is also about pushing the boundaries of technology and improving security.”
Artificial Intelligence is employed to find “fuzzy” patterns insufficient to be exact copies of design elements. The framework will eventually reach a range of results where positive matches can be confidently identified. Law enforcement provided known counterfeits to be studied using state-of-the-art scanners. Scanned images are separated into regions of interest due to what in the paper are called “blobs.” Blobs are identified based on visual similarity and composition. Blobs are then treated as clues when studying scanned images. Patterns capture relationships among the blobs’ attributes, including color, texture, shape, and size. The patterns help researchers understand the images and determine the authenticity of the coin.
The blobs play a role in generating fuzzy association rules. These rules represent implicit knowledge and relationships within a set of images, aiding in image classification tasks. The researchers hope this technique can be applied to detect not only counterfeit coins but other objects as well.
Suen said, “This method can be used to detect all kinds of fake goods, which we are seeing all over the world. It can also be used to detect fake labels on fruits, wines, liquor and so on. There are many places where this can be applied.”
Could this technology be applied not only to circulating coins but to older collector coin authentication as well?
You may also like:
