Typo spells the end of patent challenge
What happens when a patent is challenged as being obvious based on a previous patent, but the earlier patent contains a typographical error? It depends on how obvious the error is. And you might be surprised by what counts as “obvious” for the U.S. Court of Appeals for the Federal Circuit, which hears all patent case appeals.
It recently found such a patent error obvious — even though the error wasn’t discovered until an expert conducted extensive analysis 20 years after the error was first published. The result: The earlier patent didn’t disclose the subsequent invention, and the later patent wasn’t invalidated in view of the prior, errant disclosure.
The big picture
ImmerVision holds a patent related to capturing and displaying digital panoramic images. Panoramic lenses typically have linear image point distributions, meaning there’s a linear relationship between 1) the distance of an image point from the image center and 2) the corresponding angle of the object point to the image center. This allows for easy manipulation of the image but can also limit image quality.
The patented invention purports to improve resolution. It includes limitations requiring a panoramic lens with a specific nonlinear point distribution.
LG Electronics sought inter partes review (IPR) of the patent. Under IPR, the Patent Trial and Appeal Board (PTAB) can reconsider and cancel an already-issued patent based on certain types of “prior art” that made the claimed invention obvious and therefore unpatentable. In this case, the prior art was another patent, referred to as “Tada,” for wide-angle lens systems.
Tada describes four embodiments of its systems that use lenses of different thicknesses, shapes and separation distances. These “prescriptions,” or sets of optical parameters, were laid out in various tables.
Based on the prescription that corresponded to the third embodiment of the Tada system, LG’s expert testified that it had the same nonlinear image point distribution characteristics as those in ImmerVision’s invention. Tada doesn’t specifically state the image point distribution function, but the expert reconstructed the lens based on the prescription and found that the function wasn’t linear.
When ImmerVision’s expert attempted to replicate the opposing expert’s process, he found that the prescription that the expert relied on was intended to correspond to a different embodiment. After extensive investigation, it became clear that the discrepancy was because of a cut-and-paste transcription error between the Japanese priority application and the Tada reference being relied on.
The PTAB found that the prescription was an obvious error that a person of ordinary skill in the relevant art would have recognized and corrected. Because the correct prescription didn’t make the ImmerVision invention obvious, the board sided with ImmerVision. LG appealed the Board’s decision.
The court’s focus
The Federal Circuit applied the Yale standard for reviewing errors in prior art. Under the standard, where prior art includes an obvious error of a typographical or similar nature that would be apparent to one of ordinary skill in the relevant art — who would disregard the erroneous information as a misprint or substitute the correct information — the errant subject matter cannot be said to disclose a later inventive feature. Thus, the prior art can’t make a subsequent invention, which reflects the error, obvious.
The PTAB had identified several aspects of the prescription that LG’s expert relied on that would have made the error apparent to one of ordinary skill — for example, duplicative information between that and another prescription. The court therefore found the PTAB’s conclusion reasonable.
The Federal Circuit rejected LG’s argument that the error couldn’t be obvious, as it took ImmerVision’s expert 10 to 12 hours to uncover it. The court explained that the length of time and the manner in which the error is actually discovered doesn’t “diminish that there is an obvious error.” The Yale standard doesn’t impose a temporal requirement for how soon the error would be recognized.
The court also dismissed LG’s contention that the standard applies only to typographical errors, such as spelling mistakes. The distinction between a typographical error and a copy-and-paste error, it said, is a “distinction without a difference” that didn’t warrant a different outcome.
Sidebar: The dissent takes another angle
The U.S. Court of Appeals for the Federal Circuit’s decision in LG Electronics (see main article) wasn’t unanimous. While the dissenting judge agreed that the majority applied the proper standard, she didn’t find that the error at issue was only typographical or “similar in nature.”
The dissenting opinion focused on the fact that the error wasn’t discovered until an expert witness conducted a dozen hours of experimentation and calculation. It wasn’t noticed by prosecuting patent attorneys or the patent examiner, nor was it included in a certificate of corrections addressing typographical errors in the earlier patent. The Patent Trial and Appeal Board that instituted proceedings also failed to spot the error. Although it had been in publication for 20 years, the error was never corrected anywhere.
According to the dissenting judge, a typographical or similar error is apparent to the reader and can be conveniently ignored without undermining the content of the information. The error here, however, couldn’t be deemed typographical or similar in nature.