Machine Vision Patents Thrown Out

chalker writes "Cognex Corporation, the world's leading supplier of machine vision systems, announced today that the U.S. District Court in Las Vegas has ruled in favor of Cognex in its lawsuit against the Lemelson Medical, Education & Research Foundation. It held that the claims of 14 patents asserted by Lemelson are invalid and unenforceable , and not infringed by Cognex. Co-plantiffs included barcode reader manufacturers Symbol Technologies, Accu-sort Systems, and Zebra Technologies amongst others. These patents were classic "submarine" patents orginally applied for in 1954, but tied up in the patent office and changed over the next four decades to cover changes taking place in the machine vision field. Lemelson had threatened to sue numerous end-users, including Motorola and Ford, over the past two decades and had settled all of them out of court for over $1.5 billion in licensing fees. For once a judge has seen how ridiculous our patent system is."

Comments

"Submarine" Patents

[Deraj DeZine]

These patents were classic "submarine" patents orginally applied for in 1954, but tied up in the patent office and changed over the next four decades to cover changes taking place in the machine vision field

I didn't really know where the name came from until recently, so I'll comment on it (maybe it's ITFA, but I didn't RTFA). Apparently they're claled "submarine" patents because they only surface when necessary. Like when 3dfx sued nVidia and then nVidia countersued for trivial patents so that they could end up with a cross-licensing agreement.

If this is wrong, I've been misinformed, will apologize, and then hunt down whatever sick mind thought it could safely spread lies on the Internet (of all places!).

Re:"Submarine" Patents

[XaXXon]

Here's what appears to be an authoritative definition of submarine patent..

A Submarine Patent is a patent which an "inventor" files on a device or technology that doesn't exist yet, or which has not yet been successfully implemented. Using various procedural mechanisms, the filer intentionally delays issue of the patent, sometimes for years, until a practical implementation of the device/technology appears on the market. At that time, the filer allows the patent to "come to the surface" and demands royalties from the party who did the real work.

http://c2.com/cgi/wiki?SubmarinePatent

Not the end of the road, but a start

[Nakito]

This is only a District Court decision (District of Nevada) so it is not binding precendent elsewhere. But it will interesting to see if this is the beginning of a trend that eventually kills the major cash cow of the Lemelson foundation. The foundation's business model has been to sue everyone in sight (or at least everyone who makes image recognition systems) and then to offer a license on the patents for an amount less than the cost of defense. It's been a very effective strategy. But there are tons of other Lemelson lawsuits in the works, and I am sure the lawyers are all reading this decision very carefully tonight.

Re:Machine vision patent case

[bezuwork's friend]

The court probably got most of it's reasoning from Symbol Technologies v. Lemelson Medical, Education and Research Foundation (2002) where the Fed. Circuit held that latches applied. For those that don't know, all patent appeals go through the Fed. Circuit - it was created expressly to bring consistancy to the patent litigation field. I'd imagine that many Lemelson suits will go this way from now on, given the Fed. Circuit precedence. Actually, they have to if the same issues apply.

Lemelson built his empire on what are called submarine patents - applications which are prosecuted for many many years before issuing. Once issued, the owner then goes after the big companies who have dominated the market in the meantime. That's how he made his billions.

I used to work in the patent office. My supervisor once brought in a Lemelson application he was working on - it's original filing date was 35 years earlier. From what he said, the original application was for a memory chip, but over the years Lemelson added bits of information here and there, and by the time my supervisor got the case, it was for a microprocessor of some kind. Basically, he played the patent system pretty close to the edge, but pretty much legally, I guess.

For those who find this shocking - the current patent system measures length of patent term from the date of application, so submarine patents of the extreme nature of Lemelson's are pretty much not possible any more (except if the application is classified - a few years ago a patent came out which had been classified some 67? years previously IIRC).