Bloomberg: Sony was sued by a unit of U.S. defense contractor L-3 Communications Holdings Inc. and accused of infringing two patents for image sensors, 5,541,654 and 5,452,004, which were issued in September 1995 and July 1996, respectively.
L-3 asked for a jury trial, unspecified damages and a permanent injunction against infringing products, in an Aug. 27 lawsuit in federal court in Wilmington, Delaware. The New York-based company’s “extensive patent portfolio protects L-3’s considerable investment in its research-and- development efforts,” and Sony should pay license fees, L-3 said in court papers.
To me the both patents look similar and quite generic:
#5,541,654 Focal plane array imaging device with random access architecture
Abstract
An imaging device includes an array of plural imaging elements each of which is responsive to incident light flux to provide an output signal. Each of the imaging elements includes provision for conducting a variable time integration of incident light flux, and alternatively, also for selecting a time interval during which each of the imaging elements simultaneously conducts such a time integration of incident light flux (i.e., takes a snap shot of an image scene). The imaging device includes provision for random access of each image element or group of image elements in the array so that output signals indicative of all or of only selected parts of an imaged scene can be processed for their image information, if desired. The other parts of an imaged scene may not be considered or may be considered for their image information at a lower sampling rate than the selected parts of the scene so that image information about the selected parts of the image scene can be accessed at a much higher rate than is conventionally possible. A variable gain feature allows selective canceling of fixed-pattern noise, interference, or unwanted image information. An anti-blooming feature prevents charge from an excessively bright image source from cascading across the array. Also, a control cache memory allows control commands to be fed to the device at a high rate and to be implemented at a slower rate on a first-in, first-out basis.
#5,452,004 Focal plane array imaging device with random access architecture
Abstract
An imaging device includes an array of plural imaging elements each of which is responsive to incident light flux to provide an output signal. Each of the imaging elements includes provision for conducting a variable time integration of incident light flux, and alternatively, also for selecting a time interval during which each of the imaging elements simultaneously conducts such a time integration of incident light flux (i.e., takes a snap shot of an image scene). The imaging device includes provision for random access of each image element or group of image elements in the array so that output signals indicative of all or of only selected parts of an imaged scene can be processed for their image information, if desired. The other parts of an imaged scene may not be considered or may be considered for their image information at a lower sampling rate than the selected parts of the scene so that image information about the selected parts of the image scene can be accessed at a much higher rate than is conventionally possible. A variable gain feature allows selective canceling of fixed-pattern noise, interference, or unwanted image information. An anti-blooming feature prevents charge from an excessively bright image source from cascading across the array. Also, a control cache memory allows control commands to be fed to the device at a high rate and to be implemented at a slower rate on a first-in, first-out basis.
Thanks to J.S. for sending me the link.
I think the inventor, Peter Roberts worked at on focal plane image processing technology at Honeywell, among other things, but this is a Litton patent. Anyway, you sometimes don't know which claims are being said to be infringed upon until later but if it is related to random access of pixels, I do recall Orly's paper in particular. I am pretty sure the IR guys were playing with random access before Orly's publication. This paper also described a 3T sort of architecture with sampling cap and secondary follower for snapshot mode if I recall correctly but perhaps that was a different paper.
ReplyDeleteYadid-Pecht 0., Ginosar R. and Shacham-Diamand Y., A Random access photodiode array for intelligent image capture, IEEE Trans. Electron Devices, vol. 38(8), pp. 1772-178 1 (1991)
Random access pixel is obvious and the prior art is the relation to existing memory architectures
ReplyDeleteI do not agree with the last comment by "Anomynous". Pinned photodiodes were used for about 15 years in CCDs when Kodak got the patent for pinned photodiodes in CMOS imagers.
ReplyDeleteI think that the sue claims should be on the variable exposure time for DR entension.
ReplyDeleteFor me, the usage of pinned photodiode in CMOS is obvious ! No invention at all. The first conception of HAD invented by NEC is a real invention and revolution.
ReplyDeleteThe pinned photodiode in the CCD was a very good invention. It was preceded (although Teranishi says independently) by Hynecek's virtual phase device. In the 1960's Peter Noble suggested using a buried photodiode as a detector with MOS readout. So, if you use the argument above, it must have been obvious. But, I don't agree at all and most courts would not as well.
ReplyDeleteOne can only judge obviousness when you put yourself in the shoes of the inventor at that time. Most inventions ARE obvious to the inventor him/herself. The idea usually comes as a sort of flash of inspiration to a problem. But, would most engineers skilled in the art come to the same conclusion? Probably not, as evidenced by a lack of similar and concurrent disclosures/papers/patent apps.
Another good example is correlated double sampling (CDS), a bedrock of image sensor noise reduction. It was invented by Marvin White at Westinghouse. I think most of us would argue that it was a key invention. But, in fact CDS was being used regularly in radar signal processing at Westinghouse so perhaps it was obvious to Marvin in a flash. Still, I call it a bona-fide invention.
Agree and not agree with Eric here !
ReplyDeleteThe initial invention comes from a depth understanding of the fondamental problems. For example you have to understand that most of the dark current comes from surface state, you have to understand the surface states have a big impact on CCD operation. So all the initial inventions are NOT obvious !
But the further extension to the similar devices based on a new process is another story. A photodiode in CMOS process is not that different from the photodiode in CCD. Such kind of patent is mainly a business trick but no real scientific and technical values !
I've tried such photodiode long time ago using P+ over a Nwell photodiode, but the P+ implant had a so poor quality at that time ....
About the invention of the pinned photodiode, we can argue for the next coming 10 years I guess. But what I understood is the neither Hynecek nor Teranishi invented a pinning layer at the interface to reduce the dark current ! Hynecek used it to create a virtual phase, Teranishi wanted to reduce the lag of the photodiode. But the pinned photodiode is a great device. It realized a breakthrough in solid-state imaging, first for the CCDs, later for the CMOS sensors.
ReplyDeleteWhat kinds of patents or actual products of Sony infringed the mentioned patents of Litton? In your comment, you only explained the Litton patents and gave no evidence of Sony's infringement. Please give us the additional comments.
ReplyDelete"Pinned photodiodes were used for about 15 years in CCDs when Kodak got the patent for pinned photodiodes in CMOS imagers."
ReplyDeleteThis was an obvious extension from CCD to CMOS. The invention occurred 15 years prior. If Kodak CMOS pinned diode patent has been upheld in court, please note it. Getting licensing or cross-licensing for it does not count. Whether or not it has been upheld still does not indicate whether it is legitimate or not. As Albert said, it could be debated for 10 more years. Many of us think it is an obvious step from CCD to CMOS.
I believe it's time to rethink about "patents".
ReplyDeleteSome of them are really an advance for the scientific/engineering community, other are just a business trick.
In the latter case I believe "patents" is now a market where everyone hurry up to fill up a form and try to be the first even if it's just an idea or concept (in this case one can patent "the physics of Star Trek" and be ok for the next 10 years") or no really new value added.
I see similarities with "papers", there is a "paper" for almost everything but how many paper really add a value to the topic discussed?
How many papers are published just because some academic "metrics" are based on the # of papers? (quantity, not quality).
The 2 abstracts are identical (>90% of words) so it's really hard to understand any difference, I'm wondering how a court can judge on it if a scientific forum of 3 people involved on the topic are in contrast (and it's an helpful contrast cause it re-route the conversation to the 1st question, "what is a patent?").
thanks
ciao
Ernesto
Oh this is such baloney. I wonder what anonymous and "many of us" were doing in the 1993 time frame while Tom Lee and I decided to try a pinned photodiode (JFET photogate) for the CMOS APS pixel instead of the poly photogate. Obvious? hardly not at that time. Most of the (few) image sensor specialists in the world were busy with CCDs and snickering that I was working with CMOS, a proven disaster to date. Pinned photodiodes in CCDs used much higher voltages than CMOS used, and were out of the mainstream process flow. We weren't even sure we could get the thing to work.
ReplyDeleteThe number of image sensor specialists, and the number of people employed in image sensor technology has probably increased at least 10-fold if not more since that time. Obvious, like my foot.
(Kodak deliberately leaving JPL off the patent app probably voids the patent in any court situation, but what do I know? Esp. as JPL had already filed a NASA New Technology Report that included the Kodak guys, just like the paper we all co-authored. But, that was Kodak politics circa 1994-1995.)
Eric,
ReplyDeleteIt probably took some effort to make it work, but that doesn't make the concept of PPD on CMOS any less obvious. A patent on a specific recipe to make it work would make sense though.
Many of these anon's were still getting weekly allowances from their parents in 1993 so they didn't really get a chance to work on the obvious.
-wicky-
Well Wicky, it is all obvious 15+ years in hindsight, isn't it?
ReplyDeleteI suppose 17 years from now the same group will be pointing out in 2027 how obvious the next breakthrough was in 2010. Wait a minute...what is the obvious next generation image sensor? Tell me!
Many experts in CCD will agree that the pinned photodiode used in CMOS is an obvious extension. And those experts were designing before Fossum was managing. A fully depleted pinned photodiode with complete charge transfer is not a photogate. Fossum only pushes that concept because he thinks he is the inventor of the photogate too. The reason he pushes the photogate to pinned diode similarity is because he doesn't have any real claim to the pinned diode in CMOS, which is the obvious extension that Kodak made. Cars have wheels. You can put wheels on a truck as well. The truck is heavier than a car and we might be worried that wheels won't work on a truck. But it doesn't make it any less obvious that you can put them on a truck.
ReplyDeleteYes, the PPD in CMOS is just a simple extension from CCD. It took the processing capabilities and know-how from the CCD community to make CMOS worth looking at for mainstream products. Fossum certainly helped contribute to the community of those folks working on CMOS sensor arrays at that time. I did not realize that Fossum has a claim on the PPD. The explanation about the photogate above makes sense now.
ReplyDeleteSigh. More anonymous postings from the department of disinformation, history rewrite section.
ReplyDeleteThe similarity between photogate and pinned photodiode is obvious. Eric Fossum is right at this point !
ReplyDeleteThere is a big difference between "what I can do with this CMOS process" in the past and "what I can put better option in CMOS process" now.
If we look backward, there are a lot of trivial things. That is definitely true. But We have to distinguish some conceptual breakthough and concept extension. One is revolutionary and the other is incremental.
The pinned photo diode in CMOS is incremental. The invention occurred at the CCD. In fact, the read out circuit of the CCD is just placed in the pixel. So it is barely incremental. It is more of an arrangement.
ReplyDeleteDear Anon,
ReplyDeleteDon't agree with your comment. At that time, the MOS transistor was big and you need to be very visionary to see that the MOS transistor can become so small and so performant to invest energy in in-pixel charge transfert like Eric Fossum did at that time !!!
"very visionary to see that the MOS transistor can become so small" ... I guess you're refering to Gordon Moore.
ReplyDeleteYou may be to young to remember, but the vision was formulated in 1965, in 1993, that was OLD news.
Coming up with a recipe to make PPD work is not obvious though.
Eric, it is highly unlikely that all the anon's here have a more personally biased view on this than you have. I suggest you take some distance and give it some further thought. Acknowledging these concepts were not inventions, doesn't make your contributions any less important for the industry.
I cannot imagine you can truly believe this "5,841,126" is an "invention". As stated above, it's an arrangement, as are most of the pixel patents out there.
-wicky-
Seems to me that many of the postings are by CCD people who have steadily seen their market share vs CMOS dry up over the years. And now they are almost trying to claim that CMOS is just an obvious subset of CCD.
ReplyDelete"Obvious" is a gray slider scale - even Isaac Newton acknowledged his ideas built on the ideas of others. Determining where on the the "obvious" slider scale settings to declare something patentable is more an area of public policy than of engineering.
Wicky, this has to be your silliest post yet.
ReplyDeleteEric,
ReplyDeleteI'm amazed by the quality of your arguments. I rest my case.
-wicky-
Can someone explain the relationship between Wicky and Eric? Wicky has some really valid points and Eric seems hostile. I have noticed that Eric usually comments on IP and awards as if he thinks about them constantly. If they are so important to him, it must mean that he wishes he had invented the most significant aspect of CMOS image sensors, the pinned photodiode. Did Wicky win an award or something that Eric wanted?
ReplyDeleteNo relationship at all. I have no clue who Wicky is or what Wicky's contribution to the field might be, but I, perhaps foolishly, use my real name. And I comment often, it is true.
ReplyDeleteWith anonymous posts you never know who is posting. Certainly there is one or two anonymous posters that like to give me a hard time. Oh well.
Anyone who knows me knows that I am very supportive of Teranishi-san's invention of the pinned photodiode and especially of his recent award.
So, why don't you stop trying to be a troublemaker and let's keep the personal stuff out of this. Let's stick to facts and technical opinions.
@ "Can someone explain" and other comments
ReplyDeleteI will explain the situation for you.
Eric is a senior, respected member of the image sensor community, a position he earned with hard work and a willingness to stand up and fight when confronted with adversity.
Your post, and some others, are thinly-disguised provocations that exploit Eric's character. One guy takes an anonymous swipe at Eric, and Eric punches back. Then, with Eric in a fighting mood, another anonymous guy steps in and has a go. Before long, it's a whole group of attackers ganging up against one person.
Please stop. You know exactly what it is you are doing, and you know it is not right. Please just stop it so we can have a friendly atmosphere for discussing image sensors.
@ Ernesto
ReplyDeleteOne patent indicates it was a divisional application from the other.
This means that the patent examiner felt an original application covered two different inventions, and required the applicant to split it into two separate applications. Both applications have the original specification (the plain language discussion of the inventions) but different groups of the original claims.
@ the general readership
In US 5,452,004, both independent claims recite an "image element" comprising a "logic gate" for issuing control signals to charge manipulation circuitry. So the random access has a very specific form of an in-pixel digital decoder. Systems where instead there is a row decoder and a shared control line passed across the pixels in a row aren't covered by this patent.
In US 5,541,654, claims 1 (and 2-4) and 18 also recite this limitation. Independent claim 5 (and dependent claims 6-14) recites elements that seem to be a row address line and a column address line, but the language is vague. Power supply lines would also qualify. Claim 5 seems broad enough to cover systems with per-row control lines and no in-pixel logic, but may also make it vulnerable invalidation by prior art. Independent claim 15 and its dependents appear to have some language difficulties that I haven't had time to look at in detail (I've only had a half-hour to spend).
So my quick opinion is that the defending lawyers have some homework to do, but the litigating lawyers have weak material.
Thanks CDM.
ReplyDelete-Ernesto