Oral argument — COGNEX CORP v. ITC

On the calendar today is Appeal Number 2011-1098, COGNEX CORP v. ITC. Mr. Bauer, good morning. You may proceed. May it please the Court, there are two overarching questions that are before your panel today. First, did the Commission err when it systemically narrowed the claims of the patent to an embodiment in the patent? And second, did the Commission err when it found these two patents invalid under Section 101 Bilski? Now, with respect to the Section 101 issue, I'm not going to argue that issue today, Your Honor. I want to be clear. I know that usually when someone says, I'm not going to argue, the panel says they think this is a weak case or they don't think it's important. On the 101 issues, I think those issues are so clear, given the Old Commercial case and the Research Corp case, both of which came down after the Commission ruled, that I'm happy to submit those on the papers unless Your Honor has some questions. What's your view on the effect of the Dealertrack opinion on this case? I think Dealertrack is a very different case, Your Honor. I think Dealertrack was where they added the word computer. When you look at the claims, it was really just the computer-assisted or the computer, it was just something that was being done on a computer. But when you look at this, these claims are virtually indistinguishable from the Old Commercial case or the fuzzy, what was the recent one that just came down? The fuzzy shark case. They're just almost indistinguishable from that. This is machine vision. These are industrial processes. There's nothing about this that could be done in the human head. Why is that so? If you had somebody in business school and they had a great big class with a bunch of smart people and they sat down and said, why don't we just see if we can't map this invention out on paper? Your Honor, this is... Map it out on paper. We have our little yardsticks, our measuring sticks, we know where we'll put the probe, and they map the whole thing out on paper. Your Honor, this... Are you saying it can't be mapped out on paper? Well, I can tell you that MD Tech's expert said he couldn't conceive of doing this without a computer. And that was their technical expert. So, certainly the first person... If you agree with me that it could be mapped out on paper, and I think you have a hard time saying it couldn't, and it happened in that classroom, that's an infringing act, isn't it? It isn't, Your Honor, because this is all about what can't be done on paper unless you're going to talk about 10,000 people sitting... I mean, then theoretically anything in a computer can be done on paper. What we're talking about is machine vision. So what we're talking about is taking an image that could be thousands and thousands of bits and massive amounts of data manipulation. Massive amounts. What we're talking about here are the machines that, like at the post office, when the envelopes are zipping by so fast that you can't even see them, that it's able to scan and find... Those limitations aren't in the claim. You haven't taken a machine and then altered the machine so that it performs these steps. No, but what's in the claim is that it talks about digital images. What happens when I walk in a room and I know the portrait is not level? Why do I know that? How do you know that a portrait is not level? Yeah, if I walk in a room, you know, lots of courtrooms have portraits around and most people like to have them hung absolutely level. Right. And I see one that's not level. How do I know that? Clearly, when you see it, I mean, your brain is doing things. It's done a whole lot of things, hasn't it? Your brain has done a whole lot of things. Like in the claim. The brain isn't computing a max score. If I have two eyes, it's like a patent that has, you know, does it twice. Because if you look with one eye, sometimes it's not as clear as looking with two. Your Honor, I think that this is... I mean, when you look at the claim where it's talking about providing a model of a pattern, and it's talking about a runtime image right there, I mean, right? It's taking a predetermined pattern in a runtime image. That is a computer image generated from a camera. This is electronics. Does the claim say that? A runtime image? Does the claim say it's generated by a camera? Well, what else is a runtime image? Well, a runtime image in my mind is that when I was a little boy, and I walked in and I saw a bunch of pictures, my mother would say, well, this one's level and this one's not. And over the course of time, I got to be really expert the way a framer is. You know, people that actually hang pictures in museums, they can hang them where they don't need a level. Because they have thousands and thousands of images packed into their mind. So, Your Honor, if that's the question, whether your brain can do these things, then there's not a software patent in the world that survives that test. Not a software patent in the world, because that's what software is doing. It's running on a computer, and it's modeling. I won't say it's modeling the human brain, but the position that I think you're advocating is that anything that you could sit down and do on paper... And what's your best example of a pure software patent that we've upheld as passing 101? What case? I think Ultramershal is a very good case. Ultramershal and Fuzzy Sharp and Research Corp, I think all of these cases, they all go to digital images. And that's what this patent is all about. This patent is nothing other than digital images. And when you have their expert saying that he couldn't even conceive of doing this without a computer, I'm not quite sure where else you can go. The problem here is your 539 patent doesn't even mention a computer. It does not mention a computer, but it's the context, right? The whole patent is about a computer. And when it says, a method for determining the presence of an instance of a predetermined pattern... Well, every complex algorithm then passes 101. Excuse me, I'm sorry. Every complex algorithm. Right. Everything that has a practical matter, one would not perform in the head. So the question becomes, is this concept purely abstract, right? Isn't that the test? It's not, can you do it in your head? Although the three cases that I'm talking about all talk about whether there's a palpable technology here, whether it requires intricate and complex computer programming, which is what Ultramershal talks about. What's the palpable technology? It's machine vision. It's what? Machine vision. But that's not in the claim. Well, the claim talks about a run-time image. And run-time image in the context of this patent is an image that comes off a computer in the run-time instance. It's the only thing a run-time image means. Can you have a run-time image coming off a projector, a 35 millimeter projector? Is that a run-time image? We're talking about digital, right? Where? Well, that's what it is. Where does the claim say digital? Well, I think that was what a run-time image is. That would be a plain construction. Judge Rainer just asked you a question you didn't answer. I'm sorry, I thought I did. I'm sorry, Your Honor. No, I was asking, could I have a run-time image off of a projector? Well, what a projector generates is a form of a run-time image. Absolutely. I wouldn't need a computer for that. If it's generating a projector that's putting it up on the wall is a run-time image. Yes, Your Honor. That's right. So your argument is that in this case, however, it has to be digital. That's right. That's the distinction. It's the only thing. Otherwise, the patent describes nothing. I mean, that's what the patent is. It's a digital machine process. And the question goes back to abstract. And whether it's a claim construction, what is a run-time image? Whether that's a claim construction, right? If you just take the words in a vacuum, we can turn any software case. If the words are in a vacuum, every software case can be done in the head. Period. That's how software is done, right? People sit down and they write the code before it goes on the computer. They think about the idea before they put it on the computer. You lean heavily on Research Corp. Research Corp. has got, in its opinion, says, well, look at these claim terms. High contrast film, a film printer, a memory printer, and displayed devices. All of this structure that's in the claims in Research Corp. that the opinion relied on. And I believe in its analysis in concluding that there was sufficient machine activity in and around the algorithms to pass 101. Now, you can't point to language in your claims, I don't believe you can, comparable to the language I've just quoted from Research Corp. High contrast film, film printer, et cetera. Your Honor, I don't think this Court has ever issued a decision that says you have to have a hardware component in the process claim. This is a process claim. There are lots of process claims, or half the types of claims that are out there. There's hardware claims, there's process claims. A process by its nature is a series of steps. A process on computer processing is a... that there was hardware, indeed a computer, in the claim. As the District Court, if I'm not mistaken, in Dealertrack, construed the preamble to insert the computer into the entire patent. Upon appeal here, I believe the panel disagreed and said this is a computerless claim. That's right. Because you look at the whole... that was a business method claim. What this Court said is that was an abstract. What I'm trying to get at is you have a... on the face of it, have a computerless claim. On the face of it? On the face of your patent, you have a computerless claim. But, Your Honor, if we go there... Do you agree with that? On the face of the claim, just looking at the words, the word computer is not in there. You have maybe one peg to hang on to in the spec where you refer to some aspect of a computer, right? An image being digital. The image is digital, and that's what the patent tells you. That's the only hook in the claims, right? In the claim, that is the only hook to the computer. In the claim, in the literal words of the claim. That seems to be quite different from the halftone image application that was claimed in Research Corps. Is it not? I don't think so, Your Honor. And I think if where you're going is if the word's not here, I think then you're going to a decision that says any claim relating to software has to have the word computer in it. And I think that's a formalistic decision. Well, we have to evaluate the patent eligibility based on what it is that you're claiming. Based on what it is we're claiming. So we can't ignore the fact that the claim doesn't specify a computer or a particular, to use the Research Corps language, palpable application? Except that when you understand the claim in the context of the patent, this is not an abstract claim. So that's what we always go back to. It's machine transformation and all that. The question under 101 is, is it abstract? Is it abstract? And the question whether it's abstract is, is this something that is blocking the whole world? Or is it something that has a technical, feasible, palpable, functional aspect? Does it relate to machine vision, not human vision? And when you talk about, you're right, it doesn't say on a computer doing this method. It doesn't say that. But it says it's a method for determining the presence of a run-time image. A run-time image has no meaning other than in this context, in this context, in this patent. It's a digital image. So putting it on the wall, we can have a conversation. Is that a run-time image if I'm playing it and you can see it in your eyes? But the rest of the context, nothing then fits. Because if you're watching it on the wall, then you're not doing all these other things, providing a model that represents the pattern, right? This is not an abstract idea. It doesn't have the word computer in it. All right, Mr. Brown. We understand your position. I realize you're now in your rebuttal period, but I'm going to give you some extra time, and I want to give you an opportunity to talk about the claim construction and the non-infringement. Before you get into that, though, I have one question. We are reviewing a decision from the ITC as to whether there was or was not an unfair act under 337. If we affirm on the non-infringement aspect of this, do we need to address the 101 issue at all? You're not obligated to address the 101 issue. That's in your discretion. The cases that I've seen from this court where you've deferred the validity issue, as far as I know, I can't say exhaustively, but the ones I've seen you deferred, they were enablement, they were obviousness, they went to the very specific claim. In this case, I think the 101 issue is fundamental. This was the ITC's, I believe, their first 101 decision. This is not a cardinal chemical case where we're obligated to review the validity determination. I don't know that you're obligated. I'm not familiar with that. I can't speak to the cardinal. The Supreme Court said that when you have a case that's coming up out of a district court and you have a situation such as this, you're obligated to decide the validity issue, even if you decide non-infringement. Right. And this court has taken the say it's different. The question the presiding judge was asking, I believe, is whether or not that being the case with regard to the burden on us as a matter of law and appeal from a district court case, are we similarly burdened when the case is coming from the International Trade Commission? I'm sorry. I thought I answered that. You're not obligated. I thought I did answer that. You're not obligated. It's within your discretion. I'm sorry. Why are we not obligated? Because the issue from the ITC. What's the rationale? I'm sorry. What's the rationale? Because the ITC, the validity issue, is an affirmative defense, and it's not a counterclaim. And I can give you the case sites. There are a number of cases from this court that deal with that, that distinguish an ITC appeal from a district court appeal. And I think those cases are, if Your Honor would like those sites, I can hand them to you. But you're not obligated. Submit those afterwards. That would be fine. Okay. But I do think in your discretion, because this is a 101 case and it's the ITC's first, that you just shouldn't leave out there a decision from the ITC that says this kind of patent isn't patentable, software patent. But it's also binding on your client, right? Well, ITC decisions don't have the collateral estoppel effect of a judicial decision. It might be binding if we went back to the ITC. All right. Why don't you spend just a few minutes briefly talking about the non-infringement claim construction issues. And, Your Honor, on that, because it appears when you read the brief there's a lot of issues. I can summarize it as a single issue and really make it simple. It's a Thorner issue. The Thorner case that just came down last week from this court is exactly the problem that happened here. So we point out to a number of issues, but these are not claim construction issues where you have experts disagreeing what is an asset or what is a computer network or where you go into computer terms or anything like that. Every single issue that we point to is ordinary claim construction where the court has taken an embodiment from the patent and put it in. So what we're talking about is whether each, the ALJ talks about each being every, but when you read the claim each is referring to the element, it's each of a plurality of poses, not each of everything. But isn't it every one of that plurality? Every one of that plurality. The judge read it to be every one possible, all possible poses. He doesn't refer to each of them if you read his decision. He's not referring to each of the plurality. He says every pose. They find A, the word A, for match score surface. He construes that to be every possible match score surface. Every possible instead of the word A. For the accept threshold, which was AN, AN, an accept threshold, he says that there's more than one accept threshold and therefore there's no infringement. For test, for test, the claim has the word test. It's a broad term. He construed the test to be a test with a rating factor function. That never appears in the patent. And the same thing, and most important, is probe. Probe relates to a probe. According to the claim and according to the definition, a probe is something, and by the way, probe is a coined term. Everybody agreed. He took the construction that we proffered that came right out of the claim. A probe is something where there is a test. A probe is where there is a test and contributes evidence. What about locating and comparing? Locating and comparing. He put the word thereafter into that. So again, this is all just common claim construction rules. He put the word thereafter, saying so first you locate, thereafter you compare. And this court's law is crystal clear. Unless it's absolutely clear that there is an order required, you don't do that. And in fact, the patent doesn't tell you to do that. The patent actually, Your Honor, shows in the drawing the other way. The figure says the other way. But how is it that you can make a comparison without knowing what the item is? So, Your Honor, if you look at the claim. So here's, we're talking about a mathematical construct here. We're talking about how do you do this in a computer. And here's the way to understand it, and I'm going to directly answer that. What we're looking for is local maxima. And first, I've got to point out local maxima is plural. We're talking about a number of local maxima. Here's the easy way. You've got the Himalayan mountains. You're looking for the highest peak. And this is how you can do it either way. You're looking for the highest peak. There's two ways you can do it. One is you can locate every peak in the Himalayas. Locate every peak and get the height. Measure the height of every peak. And then you compare the height of every peak with an accept threshold. That's one way. The other way you can do it, say there's a cloud layer at 18,000 feet, so that half the peaks are under it. Now you only need to look for the peaks above the cloud layer. That's the other alternative. That's comparing the magnitude of each local maxima with an accept threshold. You're only looking now at the ones above the cloud. And then you're locating the local maxima in the surface. So the second element there is a filtering element. And in a computer, and the patent shows you that it can be both. So if you look at the patent on that element, it's Figure 22A of the patent. Figure 22A of the patent. There's the diamond in the middle where it says, Is score above the threshold and a peak? That's opposite what the claim says. So notice, by the way, it's showing them both in the same box. It doesn't matter to this operation which one you do first. It's being done in the same box. And that order on that claim, in that drawing, is the opposite. Your Honor, all I'm saying is the claim doesn't require that. And there was no reason to put the word thereafter into it. And again, it goes back to Thorner, which says you don't add words based on one embodiment. Every instance, every single instance, the judge said, Here is one sentence in the patent, or one description of an embodiment, and I'm going to impose that limitation. And so, Your Honors, I know my time's up, and I really appreciate it. I really appreciate it. I think it's so Thorner-like, and if you look at everything he's done, that I think it's easy to go through each of these claim elements. I know there's a lot. There's half a dozen of them. But they're all ordinary words, and I think it would be easy to say, Why did he say each means all this? Another alternative, Your Honor, I think is to vacate and remand in view of Thorner and say, Thorner, which just came down ten days ago, and is so clear on what he should have done. Another, possibly, would be to vacate and remand it and ask him to look at it. That's just a suggestion to make it a little easier. Thank you. Thank you very much. Thank you for the explanation. You're welcome. We'll return a couple of minutes for rebuttal. So, let's see. Mr. Jardine, you'll start. You've designated eight minutes, but I'm going to give you 12 minutes to try to even the playing field, and I will give your co-counsel also another minute or two. So, you may proceed. May it please the Court. The Commission determined no violation in this case based on nine separate different grounds. Affirmance by the Court on any one of these grounds will affirm its ultimate determination of no violation of Section 337. Not mentioned by the opposing counsel, there is a purely substantial evidence issue of non-infringement of the accept threshold limitation based on an undisputed claim construction. There are seven separate claim construction issues, and, of course, the 101 issue. Just to address one of your questions you had, Your Honor, cardinal chemical does not apply since it's not the mandate of the Commission to determine validity just to determine whether it's a violation of Section 337. What's the effect of our leaving the Commission's determination on 101 if we don't address it? In other words, if we affirm based on the non-infringement aspect of it and we don't address 101, what effect, if any, is the Commission's determination on 101? It just will remain as is. It's just the Commission determination. The patentee can't enforce the patent in the Commission against somebody else, can they? Presumably there would be a collateral estoppel at the Commission, but there would be no presidential effect going forward. But it would be collateral estoppel before the Commission. Presumably, if the collateral estoppel elements are all met, then it would be collateral estoppel at the Commission. It also has, at least, the determination could be introduced in a district court proceeding for whatever it's worth, even though it's not binding, correct? Correct, yeah. So what would be the effect on future 337 cases? Let's say there's additional imports from a different manufacturer involving the same program. You mean the same? The same invention. Right, again, if all the collateral estoppel elements are met, then presumably there'd be a... They'd be non-suited. Right, there'd be a good argument for collateral estoppel in that case. The ITC is no longer an effective place of remedy for this particular patentee for this patent. Right, assuming all the collateral estoppel elements are met, yes. But again, there's eight other non-101 issues and the court may affirm any one of these other issues. Would that be true also if, under the determination of 101, let's say that there's a determination on penability of the invention, and we say that it's not penable, would that have a collateral estoppel effect? Would you have to argue collateral estoppel in the future? Yes, it would have to be argued, of course. Assuming all the elements are met, then it would be collateral estoppel towards the same invention, same patent and claims that are found to be invalid here. Why is this case different from Research Corporation? I think, Your Honor, as you brought up, obviously these are close issues with this court and other courts, but as Your Honor brought up, the other tracts, well, Supreme Court precedent of Benson, says just a computer implementation is not enough. Your Honor has already brought up, there's a series of machines that are discussed in the background, CT scanners, MRIs, X-ray scanners, and still under Benson and Dealer tract, even if it's a computer implementation, you have to specify the how, the extent, and the significance of the computer implementation, and that's just not part of the claim here. But, I mean, certainly this is a real-world application of something of consequence, which sounds very much like Research Corporation. It's a real-world, practical, palpable application that has some practical implications. We admit there are some factual similarities, but, again, we feel that Dealer tract... Well, assume, for purposes of argument, we took Dealer tract off of the plate. Let's assume Dealer tract hadn't been decided, and Research Corp. had. Then what would be your position as to the patentability of the claims in suit in the light of Research? Again, Research Corp. was decided after the commission determination. In view of Research Corp., again, these are close cases, and there are some factual similarities. We still feel there is distinguishing elements with Research Corp., again, that even from Benson, the specificity of the computer implementation has to be met here. It's not met in this case. There is no specificity as to how... Well, how did Research Corp. get decided in the light of Benson? It's there, right? Right. Are you telling me that Research Corp. was incorrectly decided? No, no. Well, then, I still... I'm not getting the crisp answer I want from you about how we decide this case if all we had was Research Corp. and not Dealer tract. Again, we admit these are all close cases. There are factual similarities with Research Corp. and we still feel there is... there's just not enough here in the claim to tie it to a specific practical application in light of Research Corp. and all the other previous cases before. Your position is that this is so manifestly abstract that it fails to pass 101 muster. I guess using that language of Research Corp. perhaps we wouldn't use that particular language, but that's just insufficient, again, to tie it to a particular application. The algorithm pretty much just preempts the claim and there'd be foreclosing innovation in this area. Does it strike you as a little strange to conclude that the claim is so manifestly abstract that it is not patent eligible, but yet of the seven limitations in the claim, five of them are not infringed? I mean, how could it be so abstract that it's not patent eligible, but yet there are five specific reasons as to why there's no infringement? Right. Again, these different areas require different questions. 101 requires different questions from claim construction from... I mean, it's hard to say if the claim forecloses... preempts an entire field or forecloses innovation when there are five reasons why the accused structure doesn't infringe. Right. I think it's just different questions, and again, these are all close cases in this area of law, but it's just not enough here. Again, we decided before Research Corp. So you keep saying that the Commission decided before Research Corp. Assume that Research Corp. had been decided before the Commission made its decision. Would it have decided differently? It's an interesting hypothetical, again, depending on what evidence was brought forth and how it was decided. What's the evidence of this case? Again, I think we still feel that it would be distinguishable from Research Corp. It's just not enough there to tie it to a practical, specific, tangible application in view of Research Corp. and the cases before Research Corp. What does it mean to tie something to a practical application? Just some combination of steps that either you tie it to a particular application, machine. Again, there's a lot of image-capturing devices discussed in the spec, but it doesn't tie it to a particular image-capturing device or a group of them, or there's no combination of steps that are a transformative step here, some type of physical step to tie it to a practical application, even in light of a computer implementation. Well, there are specific steps that are recited, and Mr. Bauer argues that at least the word image brings in the fact that this is performed digitally, which certainly implicates a computer. Right. Again, I would point to the language of Benson, which simply says digital computer implementation is not enough. It would preempt the claim, and it just has to tie it more particularly and specifically to a particular application. It's just not sufficient to tie it. Well, what about in the spec itself and the definition of image? This is one of those nice patents that actually defines some terms. I'm looking now at A244 columns 3 and 4 in the past, and they're talking about an image that's generated by any data-processing device. Again, I would say this language is not in the claim, and if you go back to column 1, it mentions a whole bunch of image-capturing devices. Do we have a rule that says language has to be in a claim? I mean, Research Corp. says we can use dependent claims. We can find structure in dependent claims to sustain patentability. So I was just wondering, the argument that your adversary is making is that, come on, let's get in the real world. This thing is computer-driven. Again, I would admit it's a close call. Again, the court did not reach this issue. What am I to make of that reference in the spec to the fact that the image is generated by a data-processing device? Sorry, which? You're in column 3, you said? Column 4, line 5. Yeah, again, it's a close issue, but again, in language, it's not in the claim. Well, how close does it make the issue? What we've got here, we've got a case that's somewhere between Research Corp. and Dior Track. I mean, assuming that we're not going to say all these steps can easily be performed in a model mind, so we're in between. So what are we closer to? I think we're closer to Dior Track. I mean, Dior Track even had the recitation of computer-aided. That's not even in this claim here, just mere data-gathering steps, computing, comparing, providing. There's nothing tying it to a particular specific tangible application in view of all the case law. Again, it's a close issue. In my mind, I'm looking at it as we've got this, these two poles of a magnet in the form of Research and Dior Track on, and this case is being attracted to one or the other. And the question is, what force is pushing it more towards one than to the other, and what analytic tools do we use to decide that pushing and shoving? There certainly is more structure described, at least in the dependent claims in Research Corp. than there is here. I think you just have to look at the plain language. Is there any tie to a specific application? Is there any combination of steps that recites a transformative step, as you have in other cases? I think the answer would be no. There is no combination of steps that are transformative. It's just computing, providing, comparing. There's nothing determining, nothing tying. How about the plurality of probes and how the probes change and represent relative positions? And the running of tests and the resulting images at given poses. It's in there, but again, it doesn't tie to what are you doing. Are you determining something and tying it to a particular device or some transformative step? Assume that we conclude that the definition of image as being something digital is sufficient to implicate the computer implementation of this method. So then you have the recitation of all of these different steps and the specificity of all of these different steps arguably being accomplished on a computer to do something that may or may not be capable of implementation in any other way. Is that enough to render it patent eligible under 101? Again, these are close calls, but again, dealer tribe. I know you keep saying it's a close call and I think we appreciate that. We're looking for some help as to where we draw the line and how we make the decision. Yeah, I think you look at the claim language and the spec as a whole and there's just nothing tying it to some particular tangible application. You're not displaying pixels on a display. If we see the computer technology, I mean, the language in Research Corp was a functional and palpable application in the field of computer technology. Would you agree that the limitations in this claim when the method is performed produce a functional and palpable something? I mean, it forms something, but I think, again, it would foreclose any innovation in this area. You couldn't do anything with this algorithm anymore with any device. You have a particular application towards a camera or towards an x-ray device, this would foreclose that. It doesn't foreclose the accused because the accused arguably does not infringe for at least five reasons. True, but... It's a little odd to argue, well, it forecloses all innovation, it preempts the entire field, except the accused structure is okay. Yeah, I think it's all just different questions and different decisions regarding the patentability and plain construction infringement. Again, there's nine, eight other issues besides the 101. The court may affirm and may not reach this issue, but we do understand it might be the court's discretion to strongly consider this issue. All right, we've used up a lot of time. We appreciate your argument. Mr. Lurie? Thank you, Your Honor. And we'll give you seven minutes, which is what you had originally requested. If necessary, we can give you a little more time, but we're already running over. Okay, thank you, Your Honor. Matthew Lowry, may it please the Court, on behalf of NV Tech, which is the intervener in the appeal and the respondent in the ITC case. I guess I'd like to at least briefly address the Section 101 question. I think Fuzzy Sharp should probably be taken off the table as not published and actually not deciding the issue but remanding it. In connection with the Research Corporation, I think it's easily distinguished. Research Corporation, at the district court level, the dependent claims were found to be patentable because they actually effected a transformation, which is to say they converted one thing into something else in the physical world that was being displayed or printed out. And the RCT Court did reverse on the broader claim, essentially saying we're looking at this and it's all about the rendering. So if it wasn't actually saying it's a physical transformation, it was this close to a physical transformation. In the present case, there is nothing ever that happens in the real world in this claim. And to the extent that RCT Court could be read as saying we go beyond the claim and decide the broad claims and why the dependent claims, I don't think it holds that. If it did hold that, that would be error because it's inconsistent with the Supreme Court cases. What about Mr. Bower's argument that the image and the runtime image at least implicates that this is digital? There's two issues with that, Your Honor. First of all, it's not at all clear from the definition in the 539 patent that it is in fact digital and or tied to the real world. If we look at the definition that the court was citing, it defines image as a two-dimensional function, which is an interesting thing, whose values correspond to physical characteristics. Okay, that's one thing. Or whose values correspond to simulated characteristics. So it may actually be nothing. So I actually look at this and I used to play with my kids this Where's Waldo game. It's a big book and there's a little teeny figure. You've got to find Waldo in the picture. You could use this to find Waldo in that particular book. There's nothing in the claim that prevents you from doing that. You could have a simulation being done at college that actually never touches the real world. There's nothing in the claim that says that that's not infringement. So all of those things that are not done in the real world that are purely abstract are all covered by this claim because of the way that the claim is written. Whether it's on a computer or not. The definition of image does talk about a function corresponding to physical characteristics of an object. The object is defined as something measured by an image-forming device or simulated by a data-processing device. Yes, Your Honor, but then it goes on to say or whose values correspond to simulated characteristics of an object. So it could be, for example, a Where's Waldo. The RTC found that it also, some of the claimed algorithms included satellite imaging for a variety of military and civilian scientific purposes. Wouldn't that necessarily implicate digital imaging? It could, Your Honor. I don't think that tying this to digital imagery is going to actually render this patentable. If so, Benson was wrongly decided because the claim actually recited the shift register there. I don't think that the RCT functional and palpable test could possibly on its own be the test because in Fluke we were talking about an alarm limit that would keep a petrochemical plant, I guess, from exploding. I don't know if it's exploding or just entering into something. It doesn't really get more functional and palpable than that. That was described in the Fluke patent specification. The claim just said updating the alarm limit. So I don't think RCT can be read as broadly as functional and palpable in some ways enough. This is even different than RCT, though, and that's actually, I think, very important to the distinction. Because in RCT there was one specific application that met the transformation test, printing it out slash displaying it, at least according to the district court. Here, there's nothing. The only thing that's actually in the specification appears in the first two paragraphs on the background section, which under Fluke, if you're only looking at what's making it patentable and you set the abstract aside, there's nothing left but the background section. And even that says pattern location can be important, although it doesn't say how, in industrial automation, semiconductor manufacturing, electronics assembly, pharmaceuticals, who knows how, food processing, who knows how, consumers' goods manufacturing, who knows how, and many others. So there's nothing constraining what this algorithm is applied to. If this claim is so broad and so abstract as to totally preempt the field, how is it that your client doesn't infringe for five different reasons? Actually, Your Honor, I would think it's closer to eight, but let me answer the question. I won't hold you to five. Thank you, Your Honor. If one looks at the Bilski case, there's a very clear answer. In the Bilski case, there's broader claims, sure, that talk about hedging risk. There are dependent claims that include specific equations for how that's done. You don't have to use that equation. There's even a dependent claim that says, well, we'll figure out the effect of weather, or a different dependent claim, we'll figure out the effect of energy pricing on this investment using a Monte Carlo simulation over years of data. A Monte Carlo simulation is a simulation where you take a very large number of random samples, you look at them all, and then you kind of average out what happens. That's Monte Carlo, you keep rolling the dice, right? You have to have a computer to do that, frankly, for a problem set of that size. Those dependent claims were addressed in the M-Bank decision, which found them all unpatentable, and the Supreme Court decision, which affirmed that they are all unpatentable. So including additional, when you've got something that's really abstract like this, including additional levels of detail about the calculation is not sufficient to take it out of that layer of abstract. If those additional levels of detail are simply more math. But it does seem a little strange, does it not? That you say, well, it's so abstract that it preempts an entire field of endeavor. It's not that it's so abstract that it affects, it's simply not patentable subject matter. It affects the entire field of endeavor in that you can't use this math in this field of endeavor, just as you couldn't use the Monte Carlo simulation or the particular formula for price in Bilski in this field of endeavor. You're taking a thing out of this field of endeavor, which is not in the category of patentable subject matter. But that happens with every patent and every claim. You're preempting what the claim covers. In an appropriate way, where it is patentable subject matter. So if you're actually doing a chemical process, yes, you're preempting something that is patentable subject matter. If what you're doing is preempting the math, then it's not. And the opening brief of the appellant is really interesting in this regard. On page 44, complaining about claim construction, they write, under MD Tech's expert's guidance, the commission rewrote the claim so grossly that they only cover one mathematical model described in the patent. Rather, the invention as a whole, which was implemented with a particular algorithm, neither of which is patentable subject matter. So what actually they're conceding here is, you gave us only one mathematical model. True. MD Tech doesn't use that mathematical model. They use a different one. But that doesn't mean that they are, therefore, allowed to preempt that mathematical model. All right. Any last comment? If I may address the infringement issues, Your Honor. There are, I think, eight. And just briefly, I won't address them all. Well, I'd be happy to address them all, Your Honor. But I think that you don't want me to. There are two for which there is no claim construction question. Accept threshold, where the parties agreed on the claim construction, and the administrative law judge, after, I don't know, seven or eight days, full days of trial, made fact finds that says, that's not the way the MD Tech software works. The second is whether each that meets the accept threshold is returned. That's another where there is no claim construction. A pure factual finding. And each that meets the accept threshold, in fact, is not returned by the MD Tech software. They argue, well, everyone that you return meets a threshold, maybe, but actually that's not even proved. But even if it were true, each that meets the threshold is plainly not returned. They do make some argument about what the specification discloses. Their expert, I mean, the briefing is replete with arguments that weren't made below, that their expert couldn't explain or adopt. The remaining ones, I think, are pretty squarely addressed in the briefing. They are claim construction ones. The ordering of the method. You need to locate the maxima before you do the comparison with the accept threshold and the maxima, kind of the ordering thing. Essentially, what Cognix argues here is, finding the tallest trees and then cutting down the tallest trees is met by bush cutting the entire forest, because you happen to have cut down every tree. And that's plainly not what this patent is talking about. It plainly says you find them before you cut them down, because it doesn't make sense in the context of this patent to do it any other way. All right. Thank you very much. Thank you, Your Honor. Mr. Bauer, we will give you three minutes for rebuttal. That will be plenty of time, Your Honor. So first, you asked about the ITC, I mean, this Court's cases on the discretion on the... So I have two, Your Honor. It's SINORGCHEM, S-I-N-O-R-G-C-H-E-M, versus ITC, 511, F-3-R-D-11-32, which is 2007. And Solomon Technologies versus ITC, which is 524, F-3-R-D-13-10, FEDSER 2008. And I just think the difference here, those cases do say it's in your discretion not to do it, but those cases are, I think, if you look at them, very specific about those claims, and you just didn't want to spend the time, because you didn't need to, I think this one-on-one issue is much more important. Let me get right to... So first, image. We asked the Commission to construe the term image with that definition. That's the definition in the patent. It is, there's no question, it's the inventor's definition, and we asked for that definition. He declined. He gave no definition. So if that's the definition, so we haven't waived that. We asked for it. He declined. With that definition, there's no question that this thing is a computer thing. There's no question, when you talk about image-forming devices generated by data processing, Research Corp and Ultramobile, because we've been talking about them, let's look at those claims, because I may have misspoken, or at least misguided. Research Corp has nothing hardware in that claim. Nothing. It is a method for half-toning of grayscale images by using a pixel-by-pixel comparison of the image. It's the same kind of claim as this. I don't know how you distinguish Research Corp from ours if image means a computer-generated image, and it does. Same thing with Ultramershal. That claim has no hardware device in it, unless you want to say doing something over the internet turns it into a hardware device. But again, that is simply a bunch of steps, and that's in the preamble, a method for distribution of products over the internet. That's the only hardware I see in the Ultramershal. I don't see how you distinguish this claim when an image is so clearly a hardware-generated image. And dealer track, I mean, what the court was saying there is just sticking the word computer isn't what's important, because you've got to look at everything else. I've got to say, when we talk about computer, and it addresses some of the questions, is this an image if I saw it on the wall or whatever, the same argument can be made about a computer. An abacus is a computer. And so if all of a sudden you're going to take this stylistic requirement that says a computer-aided design, and then can I do it in paper? Well, an abacus is a computer by most definitions. It's not what anybody's using. It's not the word computer that makes it patentable. And saying is it a computer or not doesn't matter. This is a digital process. The word image is defined all about hardware, and that's just the way it should be read. Well, the digital image before me says your time has expired. Okay, thank you, Your Honor. I thank you very much. Thank counsel for both sides. The case is submitted.