Analysis of a 3D printer's... fingerprint? 🤨
The Atomic Layers: S10E22 (00289)
Atomic Layer of the Day:
This is certainly one of the most extraordinary uses of AI in 3D printing that I’ve come across so far!
Creating 3D models from text prompts or photos? Difficult, but fairly obvious. Managing production or a virtual warehouse of parts? That too.
But how about this — identifying a 3D printer based solely on a 3D-printed object, without knowing what printer was used?
A research team led by Professor Bill King from the University of Illinois Urbana-Champaign has developed a method to determine the origin of 3D-printed parts using artificial intelligence. This technology allows for the precise identification of the exact machine on which a given part was produced.
The research revealed that even identical 3D printers, using the same settings and materials, leave unique "fingerprints" on the parts they produce—imperceptible to the naked eye but detectable through high-resolution image analysis.
The key to this discovery lies in the fact that every 3D printer introduces microscopic variations in the surface structure of the printed parts. These subtle differences, termed "manufacturing fingerprints" by the researchers, are distinct enough to serve as identifiers for specific devices.
In their experiment, Prof. King's team analyzed images of 9,192 parts produced on 21 different machines from six manufacturers, using four different additive manufacturing technologies. Using an AI model, they were able to identify the source machine with up to 98% accuracy, even based on a 1 mm² section of a part’s surface.
This technology has broad industrial applications, particularly for monitoring suppliers and ensuring compliance with established production processes.
Prof. King points out that modern supply chains are largely built on trust, and manufacturers often lack the means to continuously verify whether suppliers are adhering to specified procedures. Changes in production processes or materials can go unnoticed for extended periods, potentially resulting in defective product batches and serious quality issues. In contrast, this new method enables quick and accurate verification of whether delivered components were manufactured according to the original order.
Notably, the system requires relatively little input data to be effective. Tests have shown that as few as ten part samples from a given machine are enough for the algorithm to learn to recognize its unique “fingerprint.”
The technology also holds potential for tracing the origin of illegal or counterfeit goods.
The research by Prof. King's team has been published in the journal Advanced Manufacturing. The authors claim that these manufacturing fingerprints have been hidden in millions of additively manufactured parts for years, and only now—thanks to artificial intelligence—can they be practically harnessed.
Atomic Layer from the Past:
05-22-2012: Objet released Objet30 Pro 3D printer.
05-22-2013: Zortrax launched its Kickstarter campaign for M200 3D printer.
News & Gossip:
Israeli foodtech firm SavorEat is entering the U.S. market with its 3D-printing “Robot-Chef” for commercial kitchens. A new partnership with a U.S. consultancy will support sales and operations from a Chicago demo center. This move marks a major step in SavorEat’s global expansion and commercialization strategy.
Swedish Freemelt has partnered with Chinese firm Jiuli to expand sales in China, Taiwan, and Hong Kong. The move supports Freemelt’s strategy to grow in Asia’s fast-rising metal AM market. Jiuli will represent Freemelt’s E-PBF technology, targeting sectors like medical and energy industries.
Lithoz has received ISO 13485 certification for its quality management system, aligning with international standards for medical and dental device manufacturing. This follows its earlier ISO 9001 certification. The company aims to support the development and production of medical applications using ceramic 3D printing, including bioresorbable implants and dental components.