Envision a container filled with tens of thousands of metal screws made from the same mold. Is it possible to tell one from another?
Yes, with NEC Corp.’s “fingerprint of things” technology — a new development in the field of forensic authentication.
“Our company has been developing authentication technologies (to recognize) facial and fingerprint (features), aiming to become the top firm in this field,” said Rui Ishiyama, principal researcher at NEC’s information and media processing laboratory. “Now we’ve turned our attention to such technology for objects.”
Tokyo-based NEC claimed in November that it had developed the world’s first technology that can identify common industrial parts like bolts and zippers through microscopic surface patterns.
Although mass-produced products look smooth and identical to the naked eye, they are different under the microscope because even items cast from the same molds bear unique surface patterns.
NEC’s technology stores those patterns in a database so industrial items can be identified quickly from enlarged smartphone photos. This can help determine where and when an item was made, expediting monitoring and tracing without the need for barcodes, serial numbers or other physical tags.
“Industrial products look very plain, but we can use fingerprint of things technology (to tell them apart). It’s similar to identical twins, who have different fingerprints,” Ishiyama said.
It might sound confusing to state that inanimate objects have their own “fingerprints,” but the technology appears quite simple.
In a demonstration, NEC showed that nearly all of the recognition tasks can be performed on a smartphone that has a magnifier attached to take big photos.
Ishiyama picked up several bolts at random from a bowl filled with hundreds of them. Each bolt has its own number tag, and as soon as he took a photo of its surface, the smartphone sent its microscopic surface pattern to a database and received its number within a second.
He repeated the ID process with several bolts and got them all right.
“Smartphone cameras are actually as good as microscopes, so if we use them wisely, it is easy to take pictures of surface patterns,” Ishiyama said, adding that fingerprint of things technology has several merits.
First, it can keep track of products people purchase and thwart the use of fakes.
For example, if a brand new bag gets damaged, the buyer can take it to the company’s official shop for repairs and prove it’s authentic and thus under warranty through a zipper photo, if the bag’s zipper and its microscopic pattern are stored in NEC’s database.
Second, it can be used to monitor industrial parts for maintenance and investigative purposes.
For example, if a bolt fails on an airplane, fingerprint of things technology can be used to check which bolt was used in what location, and when it was last tightened by the maintenance crew.
It is widely known that manufacturers buy parts in bulk to cut costs. But it would be expensive and impractical to put barcodes or serial numbers on every single bolt and screw.
This is where NEC’s surface reading technology can come in handy, because it enables people to effectively monitor the conditions of parts devoid of such physical identifiers.
One hurdle toward making the technology more convenient to use is the task of database development. An industrial database would require recording the surface patterns of hundreds of thousands of parts.
“It is unrealistic to manually take each photo, so automated systems to do this will be needed,” Ishiyama acknowledged.
Asked if the technology can be used on any object, Ishiyama said it can handle many items with distinguishable surface patterns but that transparent objects pose a problem because their surface patterns are difficult to discern.
Also, hard objects like metal and plastic are more suited to the technology because the patterns of objects that undergo bending or warping could be hard to recognize.
NEC said several firms are already using its recognition technology on a trial basis and hopes to put it into practical use this fiscal year.
This section, appearing on the second Monday of each month, features new technologies that are still under research and development but expected to hit the market in coming years.