"Down But Not Out," Feature Article, October 2004

DOWNbut not out

Checked in the courts, Jerome Lemelson's patent machine is still churning out new claims.

by R.P. Siegel

January 26, in one of the quieter rooms in Las Vegas, one of the biggest all-time winners in the patent game finally lost to the house. Jerome Lemelson was the high-rolling inventor, holding more than 590 U.S. patents, including claims for camcorders, fax machines, cassette players, bar code scanners, and automatic teller machines.

His inventions were often so far ahead of their time that, in many cases, the technology required to build them did not yet exist. Nonetheless, by tenaciously pursuing his rights as an individual inventor he was rewarded for his technological foresight to the tune of more than $1.5 billion in licensing royalties.

According to many people who knew him, a big part of Lemelson's success was that he filed patent applications that remained pending for decades, and delay worked to his advantage.

It was the delay that caused the federal judge in Las Vegas to void several Lemelson patents involving machine vision this year. And it is the delay that has made it possible for Lemelson, now seven years dead, to have numerous applications still working their way through the patent system. Some of them date as far back as the 1950s, but they cover technologies that industry relies on today.

THE 'SUBMARINE' PATENT

According to some people who found themselves at the wrong end of it, Lemelson was making use of a strategy called the "submarine" patent.

Until recently, a U.S. patent was enforceable for 17 years from the date of issue. This was seen as a reasonable time for an inventor to reap his reward before the technology entered public domain. But there was no limit on how long a patent could remain pending, and therefore confidential.

A number of Lemelson's patents took as long as 50 years from initial application to issue. By then, some of the technologies in question had become widely commercialized by others who may have known nothing of Lemelson's work. Like the warship they're named for, submarine patents would surface, fully loaded, in the middle of a marketplace with plenty to lose.

According to Lemelson's patent lawyer, Jerry Hosier, that wasn't the result of subterfuge, but of circumstance. Hosier argues that the Patent Office held up the patents, and Lemelson was a victim of bureaucratic delay. "People were using his inventions," Hosier said, "but until those patents issued, there was nothing he could do about it."

Lemelson and Hosier had their greatest success in leveraging 14 patents that covered the areas of machine vision and bar code scanning. By the time Lemelson surfaced with his first lawsuits, these technologies had completely saturated the industrial marketplace. Just about every Fortune 500 company had not only invested heavily in these technologies, but was literally helpless without them. Retooling all of their production processes to eliminate these technologies would have threatened many companies' ability to survive.

The companies that produce this equipment, such as Cognex and Symbol Technologies, were relatively small players. If Lemelson had gone after them directly, he might have collected a few million dollars and that would have been the end of it. Instead, citing the usage clause in the patent law, he went after their customers, the end users.

"Instead of collecting a small percentage of a small number, he collected a tiny percentage of an enormous number," said Jesse Jenner, the lead attorney representing Cognex in the lawsuit that ended in January. That meant every car, every television; every widget that was made with the use of this technology in any phase of production was fair game. By keeping the royalty rates modest on the large volume of products affected, Lemelson was able to collect hundreds of millions of dollars because paying was cheaper than fighting, although some paid more than $50 million.

Given that close to a thousand companies eventually agreed to terms, the grand total exceeded $1 billion. Those companies that went to court gave up or lost, finding themselves on the wrong end of the David and Goliath story, one man against a corporate giant.

ROBIN HOOD RIDES?

Adding sympathy to Lemelson's side was his record of philanthropy. His non-profit Lemelson Foundation, now overseen by his heirs, has given away or pledged more than $87 million to date, to a wide variety of programs all sharing a common theme of encouraging innovation. Among these are the Invention for Sustainable Development Program, the Lemelson-MIT Program, and the Hands-On Science Center at the Smithsonian Institution's National Museum of American History.

Some likened him to Robin Hood, taking from the rich to feed the poor. Indeed, the Lemelson Foundation has recently announced its intention to devote half its annual budget, estimated to run about $13 million, to support sustainable projects in the developing world.

The charitable foundation is separate from the Lemelson Medical, Education & Research Foundation, Limited Partnership. The limited partnership is the entity that pursued Lemelson's claims during his lifetime and has continued to do so after his death in 1997. It was the partnership, represented by Hosier, that faced off with Cognex in January in the U.S. District Court in Las Vegas.

This was a continuation of a suit filed by Cognex in 1998 and later combined with a separate suit filed by Symbol Technologies. At issue were fundamental claims to the machine vision technology.

Judge Philip M. Pro agreed with the plaintiff that the claims of 14 patents asserted by the Lemelson Partnership were invalid and unenforceable. According to the judge, Cognex and Symbol were not in infringement.

Machine vision: Jerome Lemelson and later his estate collected royalties from the vast range of industries that use the technology. That arrangement was disrupted early this year by a judge's decision in Las Vegas.

The primary point of law was the equitable doctrine of prosecution laches, which essentially means unreasonable delays. According to the court, "Mr. Lemelson systematically extended the pendency of his applications by sitting on his rights and sequentially filing one application at a time so that he could maintain copendency while waiting for viable commercial systems to be designed and marketed."

The ruling also decreed that the claims in question were invalid for lack of enablement. That is, no one person skilled in the art could produce the device based on the teaching of the specification. Judge Pro also ruled Lemelson's patents had not been infringed.

"Taken together," said Jenner, "these two rulings assert that no one, including Lemelson himself, ever built the machine vision system or bar code scanner he licensed to a thousand companies."

What's more, machine vision products marketed by Cognex and Symbol Technologies took advantage of newer technology than the system described by Lemelson and fundamentally changed the way they performed machine vision. In Lemelson's patents, analog cameras recorded videotape images that were compared to reference images. For this to work correctly, the objects being scanned needed to be positioned so as to present the image from the same angle and at the same visual scale. The computer in this system served essentially as an optical comparator.

SEEING THINGS THE OTHERS DIDN'T

The newer systems, designed and implemented by the plaintiffs, used digital imaging that allowed the image to be rotated and scaled as required, making exact positioning unnecessary. The same arguments apply to bar code scanners that don't require special positioning or capture an image of the object.

In a telephone interview, Jerry Hosier and his associate, Louis Hoffman, described Jerome Lemelson as an idea man whose success derived from three primary qualities—an inquisitive mind, a broad range of knowledge, and an alertness for the cross-pollination of ideas. He saw connections between things that other people didn't see. For example, watching the rapid oxidation of ceramic tiles on the re-entry of an early space flight gave him an idea that eventually led to a patent on a process for semiconductor fabrication using local oxidation of silicon.

"One of the things Jerry always used to say was, 'There has to be a better way,' and more often than not, he found one," Hosier said.

Lemelson's attorneys take exception to accusations that Lemelson cynically manipulated the patent system, by deliberately causing delays, and enhancing his inventions as he went along, enriching their content with items plucked from the latest journals. "For one thing," Hosier said, "any additional claims introduced had to be supported by the original specification." Specific claims that constitute the actual intellectual property must have a basis in the broad preamble known as the specification, to which new matter cannot be added once the application has been filed.

But Lemelson wrote very broad specifications. At times, his applications were given restrictions by the patent examiner, which meant that the application contained claims to more than one invention and therefore must be divided up in multiple patents. In some cases, as many as 18 or 19 separate patents came about as the result of a single application. Each of these actions added delays, sometimes lasting years.

Diagram above is part of Lemelson's 1999 patent for a "method for printing selected indicia on objects." A similar diagram is in his 1971 patent for a "method for coding and distributing articles." The drawings below are from patent issued this year, covering flexible manufacturing systems.

Lemelson's most recent patent, 6,708,385, titled "Flexible manufacturing systems and methods," was issued on March 23 this year. The idea, first conceived in the early 1950s, was to use newly emerging imaging and data processing technologies to introduce flexibility into mass production.

The patent had been decades in prosecution because of delays caused by restrictions, continuations, and interference proceedings (which are essentially interparty contests within the patent office between patent applicants to determine the first inventor of a claimed invention). There were also transfers of assigned rights between Lemelson and Molins, a British company that had licensed the technology, to further delay the proceedings. This complex history may explain, at least in part, the reason for the patent's astounding 694 claims.

The patent shows a filing date of Dec. 31, 1990, but in the "related applications" section, there are 20 other Lemelson applications, dating back as far as July 1954, of which this is a continuation-in-part. Of the 20 applications cited, six have become patents in their own right, the rest having been abandoned.

On June 8, 1995, Congress put an end to the submarine patent with the Uruguay Round Agreements Act. It legislated that all patents would expire 20 years from the date of filing, unless the Patent Office issued an extension for extenuating circumstances. Subsequently, the American Inventor's Protection Act of 1999 decreed that all patent applications must be published within 18 months of filing. Patent applications filed before the law passed were not affected.

My own experience with Lemelson dates back to 1999, when the company I was working for was one of 87 semiconductor companies cited in a lawsuit by the Lemelson limited partnership. As an in-house intellectual property champion with strong ties to the manufacturing division, I was drafted onto the team to defend the company against this action. My job was to understand our exposure, which meant determining every assembly line that contained machine vision or bar code technology and the number of products built on those lines during the time frame in question.

This was no small task in a multinational company. Fortunately for us, at least in this regard, there were a lot of products still being built in older plants that did not use these systems. Most of the damage was limited to the newest products.

Those of us on the team who were engineers, uncovinced of infringement, felt strongly that we should throw in our lot with Cognex, which had initiated a suit against the Lemelson camp in 1998 and was looking for allies. But the lawyers pointed to others that had fought Lemelson, only to end up settling. When it became clear that our exposure was relatively small and the royalty rates "reasonable," the attorneys met with Hosier and, like hundreds of companies before them, they settled.

Judge Pro's decision may or may not be the end of Lemelson's machine vision story, but there are lots more active patents still in his portfolio, including quite a few that have yet to be issued. When I checked the U.S. Patent Office database for Lemelson's published pending applications, I found 16, including motor vehicle and aircraft control systems, building alarms, robotic assembly systems, display technology, and audio photographs.

So we may not have heard the last of Jerome Lemelson just yet.

R.P. Siegel is an author, licensed engineer, and inventor who has served as a consultant on patent infringement cases. He has been a member of ASME since 1983.