“Fingerprints” of 3D printers will help to find the origin of the weapons
As in the case of fingerprints from the person, every 3D printer leaves a unique mark in the printed product. This is the conclusion of a new study, University at Buffalo, which describes the earliest accurate method of identifying a 3D printer based on the analysis of the printed object.
For example, a method which the research team dubbed PrinTracker, can ultimately help law enforcement and intelligence agencies to trace the origin of printed on 3D-printer weapons, counterfeit products and other illicit goods. “Three-dimensional printing has many wonderful applications, but it is also a dream of manufacturers of fakes. More importantly, it has the potential of making firearms,” said the study’s lead author, Wenyao Xu (Xu Wenyao), assistant Professor of computer science and engineering in the School of engineers and applied Sciences of the University of Buffalo.
To understand the method, it is useful to know how the majority of 3D printers. As in the conventional inkjet printer head 3D printer moves back and forth during printing. Instead of ink, the nozzle melts filament (plastic thread), causing its layers until then, until it forms a three-dimensional object.
Each layer object contains a tiny pleats, usually measured less than a millimeter, and called fill patterns. Ideally, these patterns should be homogeneous. However, the model type of the printer, filament, nozzle size, cusp and other factors make their own small changes to the templates, and the result is that the object does not exactly match the design scheme. For example, the printer sent the job to create the object polumillimetrovogo the fill patterns. But in fact, the size of the templates vary by 5-10% in comparison with the task. Like the imprint of a human finger, these samples are unique and repeatable. As a result, with their help, you can find out at what a 3D printer printed model.
In order to test the technology PrinTracker, the research team created five keys for doors using 14 conventional 3D printers: FDM dozen models and four SLA solutions. Then, using the scanner, researchers created digital images of each key. They also held with the help of his algorithm, the analysis of the elements of the fill pattern.
The result is based on the database fingerprints of 14 of 3D printers the researchers were able to compare the key with the printer in 99.8 % of cases. After 10 months, they conducted a second series of tests to determine the impact any additional use of printers on the ability PrinTracker to match the objects with the 3D printer. The results were the same.
The team also conducted experiments with keys, damaged in various ways, designed to hide the origin. In these tests, the accuracy PrinTracker was 92 %. Wenhao, Xu said that the technology is similar in principle to the method of identifying the source of documents that law enforcement officers, manufacturers of printers and other organizations have used for decades.