Friday, October 28, 2011

Intellectual Ventures Sues Nikon over Image Sensor Patents

Bloomberg: As Lucas mentioned in comments, Intellectual Ventures (IV) sues Nikon over 4 patents infringement, 2 of them being image sensor related:

US6,221,686 "Method of making a semiconductor image sensor" originally filed by Motorola in 2000, issued in 2001, and
US6,979,587 "Image sensor and method for fabricating the same" originally filed by Hynix in 2002, granted in 2005

The two patents are a subset of the image sensor patents mentioned in IV lawsuit against Canon filed in court a month and a half ago.

According to the complaint filed Oct. 26 in federal court in Wilmington, Delaware, IV initially approached Nikon in 2008 about taking a license to the patents. Nikon declined to meet with the technology-licensing company at that time, according to court papers. In 2011, IV said it made a second approach to Nikon and its “good-faith efforts to negotiate a license” with the camera company failed.

The case is Intellectual Ventures I LLC v. Sendai Nikon Corp., 11-cv-01025, U.S. District Court, District of Delaware (Wilmington).


  1. Is this another example of how IV "develops technology"?

  2. Nathan Myhrvold has so fully rotted from the inside that I suspect he'd say yes.

    (Not a personal attack, simply fact.)

  3. "It's money that matters", this time not from Neil but from Randy Newman ......

  4. it seems that this kind of structure is recommended by 4T CMOS foundries. If a designer uses this structure, who will be sued, the designer or the foundry??


  5. la structure HAD de SONY has already this kind of structure. What is the fundamental difference between all these "inventions" compared to SONY HAD pixel?

  6. I think you guys miss the point. The patent exists so you can either argue that you don't infringe the patent (and they must prove you do), or you argue that the original patent is invalid. Either the original inventors cheated, or the examiner made a mistake on each and every claim and missed some prior art word for word.

    These are all difficult defenses and it is the beauty of the troll profit model that most companies would rather settle than duke it out claim by claim for each patent that IV says is infringed. An ugly beauty perhaps. Personally I don't like it one bit. But, at this time it is perfectly legal despite the fact that it goes against the original intentions of the patent system.

  7. The original intention of the patent system is to encourage companies to disclose their secrets in exchange for the legal protection from copycats. In that case the patent system has achieved its targets: the secrets were publicly disclosed and now the patent rights are being enforced.

    I think the real abuse of the patent system lays in the huge number of patents, so that there is no practical way for the company to make sure that its product is IP-clean. And the public disclosure is beyond the public's ability to digest this huge information stream.

    May be a good business opportunity here is to create an artificial intelligence engine that formally and systematically goes through the all circuits, layout, mask set, process deck, package drawing, etc. and gives a list of the patents infringed by this product. Then company can decide whether to license or to work around these patents.

  8. Vlad, you can be sure the architects of the patent system, from the Greeks to the British and Italians (1400's) to the days of Lincoln had no intention to protect the rights of patent trolls. Non-inventive, non-practicing entities were not anticipated at all.

    It is surely a contributing factor that patent examiners are ordinary humans and with a 5 year latency, interference is practically guaranteed in high tech. So, we need a kind of "Watson" for the examiners (and outsiders) to search existing patents. I actually discussed this with the USPTO executives at the NIHF induction event dinners. Unfortunately funding for such a project does not exist.

  9. I'm not sure the patent trolls role is completely negative. For example, what does Motorola do with its CIS patent after it exited the image sensor market? Naturally, sells them for the highest price. If a troll pays this price, it just helps Motorola to recover the lost money, at least, in part.

    Now, imaging an individual inventor who discovers that a big company infringes on his patent. What does he do? One of the possibilities is to get the troll with all its financial power do the costly legal work. True, the inventor gets just a fraction of reward, but it's still better than nothing that he would get otherwise.

    One of the recent IV deals was with McGill University, so that now IV represents McGill patents. I suspect that McGill decided that it's cheaper for them to make IV do the hard work on looking for those who infringe on their patents and then suing them. Now McGill delegates its rights to IV and gets some funds in return. Is this necessarily a bad thing?

    To me the trolls are a necessarily element of any thriving patent system. The system itself is obsolete and probably superficial in the modern world, but this is another story.

  10. We don't see this the same way at all. First of all, I think the basic patent system is fine and more important than ever. I just think non-inventive non-practicing entities should not be allowed to file suit. Simple. Suddenly patents are greatly devaluated, much time is saved, and the world is a better place. You are either the inventor (inventor's institution) or you buy the patent to practice the invention. You don't buy it just to sue. For Motorola, tough bananas. The "lost money" was lost when they failed to commercialize the technology. Same for Kodak. But for some of the Kodak IP, it was a practicing entity that purchased the patents.

    Anyway, feel free to post the last word. I am done with this for now.

  11. A patent grants a monopoly for a period of time to practice the patented invention. That is all. Any other use of a patent was not included in the original intent. Intellectual property separate from practice is a construct built by lawyers. We have many things we have to endure that result from constructs not intended in the original law.

    I don't blame the lawyers - they are jsut trying to build their business. I do blame the Congress for letting these constructs, including the abuse of the patent system, continue.

  12. Dave, probably the monopoly right should be complemented by the possibility of licensing. Then what happens if the company shuts down or no longer maintains the patent-related business activity? I mean cases like Magnachip, Motorola, Kodak CMOS business, etc. Should they be allowed to sell their IP?

    Eric, this is a nice idea. Doing a mental experiment how it can work, I think all the big companies end up with the broad cross-licensing agreements, while smaller guys would remain exposed, so that any one of the big can kill them with a lawsuit.

    Also, there is a lot of possibilities for trolls to work around this restriction. For example, if IV decides to sue image sensor companies, it can create a small image sensor design outfit, invest some money into it, transfer the related patents - and voila, there is a new "practicing entity" that can sue.

  13. Vlad, I still must respond.

    For your first point, I think this is already the current situation. Fortunately big companies usually only sue when little companies become big enough to become a threat. Fortunately big companies then have a lot to lose if they lose that fight. Trolls sue little companies anyway.

    For your 2nd point, it is easy to define what a practicing entity is. Judges do have discretion when allowing suits to go forward. If a troll goes to the trouble of creating a real practicing entity then yes, I believe they earn the right to sue. Then they are also vulnerable to countersuit. I don't think their shareholders will like this model as much.

    Regarding Dave's point, yes of course they can sell their IP. But the only ones that should buy it are practicing entitities.

    I think IP should generally be regarded as a defensive weapon to protect a business that is making some product.

    We need to keep creativity in technology moving forward. Non-inventive, non-practicing patent holding companies that merely try to milk the productivity of others without contributing to that technological field are abusing the spirit if not the law of the patent system.

    I think the microelectronics area is just the tip of the iceberg. Biotech will lead to great riches for these trolls down the road and likely lead to further increases in consumer health care costs.

  14. I think Eric's idea would be the best from the standpoint of the fair and equitable development of technology, but it will never happen. The universities and the associations representing small inventors would have a heart attack. Plus considering that Congress just passed a new patent bill, the odds of them doing any major tinkering in the next 20 years are slim (the last major patent bill was in the 1950's). Too many businesses these days revolve around creating IP instead of product.

  15. Eric, I agree with you that this is an endless discussion. Still, it's tempting to continue, even though it has no any practical meaning, as you said.

    Regarding the troll-established sensor design outfit, it can only be sued if it manufactures something. I meant the outfit created solely for design or research purposes, not manufacturing. To look more convincing in front of the judge, I would even take some well-known person from the industry or academy and subsidize his dream research, but with no intention to go in production. If the research brings some worthy fruits, they can be sold to other companies to deal with production - a sort of private mini-university created by a patent troll.

    To me the trolls are the innate product of the patent system. Modifying the system would change the trolls tactics, but can not put them out of business.

  16. Eric, please what is the procedure to invalide an issued patent in USA?

    -yang ni

  17. Yang Ni, I can give you a non-attorney answer but I am sure it much more complex than this.

    The "easiest" way to invalidate a patent is to show that inventors outright lied to the USPTO in preparing the application. For example, they intentionally omitted prior art that would have led to claims not being allowed. This is not so easy, really, because it is hard to prove. It is an ugly process because it calls the integrity of the inventors and their institutions into question.

    Another way is to prove the patent author was not the first to invent. This is also difficult after-the-fact unless there is a undeniable smoking gun for evidence. Also, the law changed to first to file so this won't work in the future.

    Another way is to claim the examiner made an error and overlooked prior art. Or, similarly, that another patent was being examined with an earlier priority date and therefore interferes with the patent in question. Either way you have to show word by word in each and every claim of the patent that the prior art existed. This is tedious and very hard to do since sometimes a single word can make all the difference.

    Another way is to try to invalidate a patent by saying certain things were obvious to those skilled in the art at that time. This is almost impossible to argue after the fact, especially if no one had reduced the invention to practice prior to the priority date. If it was so obvious, how come no one did it before, for example. Also you have to argue that the US Government made a mistake in each and every claim of the issued patent. Also difficult to argue.

    I guess maybe the last category is procedural error. I don't think this happens often.

    I am no expert on this. Every case I have ever been involved with as a consultant (except for current cases) has been settled one way or the other out of court. Same with cases that involved my own IP.

    Bottom line: It is difficult to invalidate a patent once it is issued.

    This discussion is just my opinion on how it might be done and has no legal standing whatsoever, and could be totally wrong. (disclaim, disclaim, disclaim!)

  18. Isn't the issue not whether IP can be sold and enforced by new owners, but rather that the licensing fees are legally set so high so that an enforcement in court would be a death sentence? Trolls can exist, but the financial gain they should expect should be trimmed to reasonable limits.

  19. This discussion thread might not solve any actual patent problems, but for me personally it is reassuring to know there are thoughtful, honorable people out there.

  20. Thanks Eric! But I think that it's also hard to demonstrate that a design violates some claims in an issued patent too, am I right?

    -yang ni

  21. Regarding what Eric said, proving that an applicant lied is something that typically can only be found out during the discovery process, meaning litigation is already in progress.

    Regarding first to invent / first to file, that generally refers to matters in an interference proceeding. An interference is when two applicants claim the same thing. That doesn't mean simply that they disclose the same general invention. That means that the actual claim language sought by both applicants has to be identical in scope. This is generally a matter encountered during prosecution of the application, as opposed to after a patent issues. Interferences are very rare and are usually intentionally caused by one of the parties.

    Regarding invalidating a patent by producing "prior art" (prior art refers to any documents or evidence that can be used against the application), the first one Eric mentions is anticipation, and the second one is obviousness. Anticipation means a single reference discloses all of a claim's limitations. If you find a prior document that clearly and unambiguously discloses every single limitation of a claim, then that is a very good way to show invalidity of a claim.

    Obviousness means that a single document discloses part of what is claimed but lacks other portions, but one or more other documents disclose those missing elements, and they could be combined with the invention of the 1st document using disclosed or known reasons. Obviousness is difficult to prove in court because the determination of what is or is not obvious is a very subjective determination and courts defer to the USPTO judgment usually.

    Another thing that companies do is to initiate a reexamination of an issued patent. They have to produce evidence (e.g. prior art references not previously considered by the USPTO) that show that one or more claims are likely invalid. I'm not sure of the details of how these are initiated exactly, but the claims are then re-examined in view of the new references and may be rejected by the examiner.

    Keep in mind that even if the person suing you has 10 claims, and you're pretty sure you can invalidate 9 of them, if he's got you on the tenth one, you're still screwed.

    *** I'm not a patent attorney, so take this with a grain of salt.

  22. Here's an interesting paper on Intellectual Ventures and the way the company operates. Kind of long, but a good read.

  23. Return of the Patent trolls part 2

  24. I think that the anticipation should be the best way to invalide a claim. Just an example, I've got the idea to share the RST and SEL of consecutive pixel lines, then I found a Kodak patent claiming the same idea so forget :( ! But later I found a paper in IEEE talking about the TFT fingerprint sensor in which they used the same idea already several years before the Kodak patent.

    Do you think that this can be considered as an anticipation againt Kodak patent ? What is your opinion, Eric and other anonymous ?

    -yang ni

  25. (1) The review paper mentioned above was a fun morning read. I disagree with the suggestion on p. 19 that Campinas University focusing on "non-patentable activites such as teaching...would be a bad result on all levels". Otherwise, it was interesting to learn what "privateering" and "defense" mean in new contexts.

    (2) @ "the best way"

    I would say "Yes!" to this and "Maybe" to the end-of-post question. You have to look at the details of the patent claims and the prior art. It's not really about the general idea being the same, it's about whether all of the specific elements recited in a given claim had already appeared in the IEEE article.

    (3) There are only 2 patents that sparked this big-picture discussion. I think ISW readers should take a look at them. Maybe they are weak and would fall apart under scrutiny from a couple dozen motivated subject matter experts.

  26. @ yang ni
    Like CDM said, it depends on Kodak's claims. The document you found has to fairly clearly show all of the elements of Kodak's claim to anticipate it.

    Keep in mind that the test for infringement is the same: for you to infringe Kodak's claim, you have to have each and every element of the claim present in your product. If you don't do or have any element of the claim, even a trivial one, you won't infringe the claim.

  27. For instance, let's take US 6,221,686.

    There are three independent claims - 1, 14, and 19. Claims 2-13 depend on 1, claims 15-18 depend on 14, and claims 20-34 depend on 19.

    The text of 14 is as follows:

    A method of forming an image sensor comprising: using an implant at a first angle to form a conducting region of the image sensor; and forming a pinned layer at least partially within the conducting region.

    There are no limitations on what the "first angle" is in this claim or its dependent claims. Likewise, while the specification gives "typically" and "preferably" recommended values (see column 3 lines 32-35) the first angle is not limited there either.

    Since there aren't any limitations to the implant angle, this claim would be anticipated by any prior art where an implant at any angle forms a conducting region and there's a pinning layer overlapping the conducting region.

    Claims 1 and 19 both recite a step of:

    forming an enhancement layer on the substrate, the enhancement layer having the first conductivity type and a first doping concentration;

    There are no limitations to the "first doping concentration" recited in claim 1, 19, or their dependent claims. The specification gives "typically" and "preferably" recommended values (see column 3 lines 2-6), but does not require that the "first doping concentration" be non-zero or different in the enhancement layer and the substrate.

    Without proper limitations recited on the "first doping concentration", prior art where there isn't formally an enhancement layer (e.g. it's just a part of the doped substrate, and they are formed together), but which does have all the other elements, anticipates these claims.

  28. Hi CDM, very instructive your comments!
    Indeed, if we follow this way, any implantation uses a slant angle. So all these doping claims have no value (so many anticipations available).

    I understand now why in the US patents, we use often "recommended, prefered" values in the claims.


  29. Ah! CDM, could you find the initial NEC patent on HAD CCD?


  30. As fun as it is to dissect patent claims, without attorneys and sometimes patent examiners involved, you can get carried away. So, please use great caution. Also, even the meanings of words can be debated. For a simple example, is 0 degrees an angle? Is 0.001 degrees an angle? More complicated: What is a pinned layer? Engineers generally understand the meaning of these things but when they must be understood by a non-technical jury for example, detailed definitions must be worked out by each side and in the case of disagreement, adjudicated by the presiding judge. It is not as black and white as it seems.

  31. The working out of definitions that Eric mentions is called "claim construction". It's where the court tries to determine the meaning of all the claim terms to determine what does or does not fall inside the claim scope. It's notoriously unpredictable.

  32. I was thinking about a Markman hearing.

  33. The text of claim 8 reads:

    The method of claim 7 wherein implanting the first dopant at the first angle includes implanting a first doping concentration substantially normal to the substrate and implanting a second doping concentration at the first anglel.

    To me, " the first angle...includes implanting...substantially normal to the substrate...and the first angle" makes logical sense only if "the first angle" and "substantially normal to the substrate" are the same angle.

    I would also find it contradictory to argue that, for this claim, implanting "at the first angle" means implanting "substantially normal to the substrate" and also at a substantially different "the first angle", while for claim 14 "an implant at a first angle" somehow - in spite of no clear limitation that I was able to find - excludes "substantially normal to the substrate".

  34. PS The word "anglel" should be "angle".

  35. Eric, you are right. This is fun!


All comments are moderated to avoid spam and personal attacks.