A team of researchers at the Nara Institute of Science and Technology (NAIST) have developed a novel technique for watermarking 3D-printed objects. With cooperation from public printing services and printer manufacturers, the researchers say using this technique to automatically mark all printed parts with the printer's serial number and date, for example, could help make 3D-printed objects more traceable. One application for such traceability could be preventing crimes such as illegally reproducing keys, counterfeiting, and even printing weapons.
Currently, there are a number of ways to 3D print a product. They include laser sintering, digital light projection, stereolithography, electron-beam additive manufacturing (EBAM) and, of course, fused filament fabrication (FFF), where layers of melted plastic are built one upon another to create a model.
The latter method is what NAIST researchers have focused on.
According to the NAIST researchers, “Generally, plastic deposition is performed with layers of constant thickness. However, in [our] method, pairs of… layers are selected, and their thickness balance is modified according to the information to be embedded.”
By varying the thickness of the layers in an FFF print, a “digital watermark” can be placed on the object, lending the piece a verifiable marker. To retrieve and verify the digital watermark, certifiers can simply scan the item in a document scanner, zoom into a high-resolution scan of the object and view the watermark.
This research is not the first to explore watermarking in 3D printing. For example, 3D printer manufacturer Rize offers full-color printing technology which can be used to mark parts with ID numbers, logos or other data. What's interesting about this paper from NAIST is that developing this method of watermarking FFF parts was an engineering challenge.
According to the IEEE paper, "A watermarking method is typically evaluated using three main characteristics: imperceptibility, robustness, and capacity. In the 3D printing context, similar to other media, imperceptibility means that modifications are not visible to the naked eye, but this definition is also extended to the preservation of the mechanical properties of the object in the CAD context. Robustness means that the watermark can resist various attacks, such as printer or scanner inaccuracies, deformation and surface degradation. Capacity refers to the number of bits of data that can be embedded in the object."
In order to develop a method of watermarking which met these characteristics and which would also work on a wide range of different FFF hardware and decode/encode reliably, the researchers focused on redundancy. "Our method meets these requirements by locally modifying the layer thickness on small patches of the surface of an object, and applying it to multiple regions of the object for redundancy," explained the IEEE paper. "Even if the object surface is locally degraded or cropped, we can still decode it if some patches are intact. Our method also provides low shape distortion, fast extraction from only one pose instead of a full scan, and blind detection. For flat surfaces, the watermark can be extracted using a standard 2D paper scanner and does not require any complicated or expensive equipment. Our method works on any fused deposition modeling (FDM) printer that provides access to the motor controls. This includes all printers with ‘G-code’ support, which is one of the most common file formats for 3D printing. We focused on FDM because it is the most frequently used 3D printing technology with a 67.7% share based on 3dHubs 2018-Q4 trends."
Aside from potential applications, this research is interesting as it required researchers to develop standard, cross-platform solutions for an industry which is largely still "the wild west" when it comes to deposition processes.
The paper is available here.