[00:00:00] Speaker 00:
Our next case is sample monitoring also versus the line 2025, 1879, and 1881. Good morning again, Mr. Sandoval.

[00:00:12] Speaker 02:
Good morning again, Your Honor. And once again, may it please the court. So just a level set here. These consolidated appeals relate to the 248 patent, which is about determining a gap between or an amplitude of spacing. I call it a gap between an aligner and a tooth using deep learning. And then a 945 patent is about determining a fit score for an aligner on a tooth using image processing, so non-deep learning.

[00:00:40] Speaker 00:
So this is all obviousness.

[00:00:43] Speaker 02:
This is all obviousness.

[00:00:46] Speaker 00:
It's the one case of this trio dealing with 945, but it also deals with 248, which was the first case.

[00:00:54] Speaker 02:
That's absolutely correct, Your Honor. And if I may, I'd like to start with an issue that cuts across both patents and both appeals, and that's the absence in the prior art, in any prior art, of a teaching of an aligner in the service position of imaging. an aligner in the service position, of capturing an image of an aligner in the service position. Now, here the board found that Kim at least suggests that a photograph may be acquired while the aligner is being worn. And if it didn't expressly teach that or suggest that, it would have been obvious in view of the teachings of Invisalign and Salah.

[00:01:31] Speaker 02:
And I'll take those in turn. But at the outset, I want to say that the board's phraseology at least suggests it is really a tacit concession that Kim lacks a teaching of imaging and aligner on teeth. And indeed, when we read Kim, we see that the entirety of this reference is about handling information on teeth and information on aligners separately. And there never is an imaging of an aligner on teeth. And it's important here to understand what Kim is.

[00:02:03] Speaker 02:
Kim is a system for assessing teeth alignment based on tooth position, tooth crowding. That's what all these patents are about.

[00:02:14] Speaker 02:
Not necessarily, Your Honor. Tooth alignment assessment is different from aligner fit assessment. Those are two very different things. One is just looking at the teeth, determining how crowded the teeth are, where the teeth are, to come up with a prescription to treat them, a clear aligner or a braces prescription to treat them, Aligner fit is a very, very different operation. Aligner fit is looking at the aligners on the teeth in the context of a technical invention using computer processing, image processing, or deep learning to look at an image of an aligner on teeth itself.

[00:02:53] Speaker 02:
and make determination. So we're talking about pure teeth assessment on the one hand and aligner fit on the other. And Kim is purely about teeth assessment. And when we look at Kim, we see that it handles information on teeth and information on the aligner in a completely separate way.

[00:03:15] Speaker 02:
So I'd encourage the panel to look, for example, at the block diagrams and flow charts from Figure 1 and Figure 6 of Kim that we produced at page 44 of our blue brief. This shows the workflow of Kim. And in that workflow, aligner information and teeth information are always handled separately. They're transmitted separately, they're received separately, and they're stored separately in separate databases. If we look beyond that, we see 20 photographs in Kim.

[00:03:47] Speaker 02:
I think a total of 23 photographs in Kim. Every single one is an image of teeth alone. None of them include the aligner. And that's because Judge Lurie and Judge Schall and Judge Chanto Kim is about teeth assessment. It's not about aligner fit assessment. And the text of Kim backs this up. The text of Kim talks about a clear aligner to be currently worn by an orthodontic patient, this future verb, to be currently worn, not being currently worn.

[00:04:19] Speaker 02:
That's an express teaching that we are talking about information here about an aligner, not information about an aligner on the teeth. Now, the board read that passage to find the suggestion of imaging an aligner on the teeth, imaging an aligner in the service position. But when we look to see what that part of the reference is talking about, we see that that's not the case. The reference is talking about information about accurate steps of a clear aligner that the doctor is providing to the patient.

[00:04:51] Speaker 02:
In Kim, the doctor and the patient are in separate locations. The patient images their teeth at home, sends that to the doctor. The doctor looks at it in a remote location and then sends back information about the next aligner or an aligner that the patient will wear. So since it's information that the doctor is sending to the patient, it can't be anything about an aligner on teeth. It can't be anything about aligner fit.

[00:05:18] Speaker 01:
And just to be clear, the paragraph that I think the board at least partly cited, the paragraph 66 of Kim at page 6301, doesn't have this language about to be currently worn. It says about a currently worn clear aligner, which is...

[00:05:42] Speaker 02:
pretty clearly ambiguous about what it's not ambiguous about is that the information is about the aligner only and not the aligner on the teeth. It's only information about the aligner.

[00:06:07] Speaker 02:
So then I'll move to the boards, what I'll call the boards. I mean, it was an obviousness finding, but I'll move to the boards finding that if Kim doesn't suggest this, it would have been obvious in view of Salah and Invisalign. And this is clearly not the case. And again, it's important to understand why. what Salah and what Invisalign are. And I'll start with Invisalign. Invisalign is a clinician's guide for conducting in-office checkups on patients who are wearing clear aligners.

[00:06:38] Speaker 02:
It has nothing to do with acquiring images at all. It's just a guide for a doctor. Nothing to do with acquiring images on teeth, nothing to do with acquiring images on aligners, nothing to do with computer analysis of images. And this is important because that is what these claims are about. These claims are about acquiring images, in this case images of a liner on teeth, and doing things with them. analyzing those images. In the case of the 248 patent, analyzing them with deep learning.

[00:07:09] Speaker 02:
In the case of the 945 patent, analyzing them with image processing. A line, the Invisalign reference is nothing like that. All it is, again, is a guide for a clinician, and all it does is show a picture of an aligner on teeth. It doesn't teach acquiring that image. It just shows the picture. And as a matter of fact, when we look at that picture, it's not even a picture of an aligner. And I'm looking at figure A now in the Invisalign reference. I believe it's on page one.

[00:07:40] Speaker 02:
It's not even an aligner in a true service position, as the claims here require, an aligner in a service position. It's a photograph of an aligner that has been purposely partially inserted to illustrate to the doctor what a large gap might look like. It's not a real-world situation that any doctor would consider at all, would see at all. So no POSA looking at that picture in Invisalign.

[00:08:09] Speaker 00:
You mean to speak English, you mean a person of skill in the art?

[00:08:12] Speaker 02:
Yes. Thank you, Your Honor. Excuse me. No person of skill in the art looking at that picture in Invisalign would have ever thought to convert Kim from a teeth alignment system to an aligner fit system and then capture an image of an aligner in the service position. There's simply nothing there that would lead a person of ordinary skill in the art to do that and certainly nothing there to teach a person of ordinary skill in the art how to do that. And Salah doesn't help aligns cause either.

[00:08:44] Speaker 02:
Salah is purely about imaging teeth. It has nothing to do with aligners. It makes no reference to aligners whatsoever. Now, the board found that Salah teaches using a cell phone to take an image of a patient's teeth. while the patient is wearing what the board called an orthodontic appliance. But this orthodontic appliance is just that retractor that I was talking about in the case of the earlier appeal, this retractor that goes into your mouth to pull back your gums, your lips and gums, so the image can be taken.

[00:09:16] Speaker 02:
It's not something that was being imaged. In Salah, they weren't trying to image the retractor. It's just a tool that helps the patients more easily take images of their teeth. So this was in no way a suggestion to image an aligner in the service position. And as a matter of fact, Salah teaches away from imaging an aligner in the service position by prioritizing high contrast areas as opposed to low contrast areas.

[00:09:45] Speaker 02:
imaging an aligner on teeth indisputably involves imaging low contrast areas, and that's because aligners are invisible. Aligners are purposely made to be invisible, and therefore they are very, very low contrast when it comes to contrast as compared to the teeth they are imaging. So at bottom, we do not believe that Kim teaches or suggests imaging of an aligner on teeth. Kim is purely about teeth assessment, and there is nothing in Invisalign, a mere dental guide, a guide for a practitioner, which just shows a picture of a misplaced aligner, and nothing in Salah, which doesn't relate to aligners at all, that would make obvious taking Kim and modifying Kim to add imaging of an aligner on teeth.

[00:10:40] Speaker 02:
there's no questions on that aspect, then I'd like to go to the 248 patent, an issue that's unique to the 248 patent, and this is the improper use of Meninas for deep learning. Meninas requires, I'm sorry, the Meninas reference for deep learning. The 248 claims require using deep learning to figure out a gap between an aligner and a tooth. Meninas teaches generic deep learning algorithms, deep learning algorithms for detecting contours in things like cats and dogs.

[00:11:12] Speaker 02:
Meninas has nothing to do with dental. Meninas has nothing to do with teeth. And deep learning devices, deep learning algorithms, deep learning databases are not fungible. You can't take a generic deep learning algorithm like the one that's disclosed in Meninas, and plug it into a dental system to determine a gap between a tooth and an aligner. Now, the claims here don't require a particular type of deep learning device, but they do require using a deep learning device to determine this gap between the aligner and the teeth.

[00:11:50] Speaker 02:
And in order to do that, you need to be able to distinguish

[00:11:55] Speaker 01:
You just said a couple of sentences ago about you can't take an off-the-shelf deep learning device and do this with it. I thought this is the same 248 patent. This is the 248.

[00:12:08] UNKNOWN:
101.

[00:12:08] Speaker 01:
Yeah. Does not the...

[00:12:11] Speaker 01:
Spec, say, here are a whole bunch of deep learning devices, quite a lot of them, actually. You can use any number of them.

[00:12:19] Speaker 02:
The spec does refer to deep learning devices, but the claims require using one to determine a gap between an aligner and a tooth. And in order to do that, and this is where I was going, Your Honor, you need to be able to distinguish between an aligner and a tooth. The technique that's used in Meninas can't do that. The technique that's used in Meninas works on a pixel basis. It's a different kind of deep learning technique than a deep learning technique that would be required to determine a gap between a tooth and an aligner.

[00:13:00] Speaker 02:
And that's supported by our expert's declaration, for example, at Appendix 63-63 to 63-65, the differences between these deep learning devices. I think I'm running out of time, so if there's no more questions, I'll sit down and reserve two minutes for rebuttal, if I may. Thank you, Your Honors.

[00:13:26] Speaker 03:
May it please the Court. Good morning again, Your Honors. So I'll start with the aligner in a service position issue.

[00:13:34] Speaker 03:
As the Board found...

[00:13:37] Speaker 03:
Kim at least suggests it teaches such a thing. Your Honor, Judge Toronto pointed out the same paragraph that I was thinking of, which is paragraph 66, which doesn't have the 2B language in it. Paragraph 66 also is what the board relied on to dismiss their argument that Kim wasn't using the image of an aligner for the purposes of clinical analysis, that it was just sort of wanting to see a picture of a clear aligner.

[00:13:59] Speaker 03:
Kim is actually using the images for the purposes of analyzing the patient and the patient status and whether or not the currently worn aligner is still the proper aligner to be worn. So Kim at least shows us all that. That's enough to support the board's obviousness conclusion, but we have more. We have Invisalign, which actually shows us a picture of an aligner in a service position.

[00:14:22] Speaker 03:
Opposing counsel referred to an image in Invisalign that may be showed in a line or not quite in a service position. I think he was maybe thinking of the image at page 6319 that shows that. But if you look on the first page of the Invisalign reference, at least the first page with images, that's at appendix page 6313.

[00:14:42] Speaker 03:
you'll see the upper left-hand picture definitely shows an aligner in a service position. And, in fact, what Invisalign wants the clinician to look at at that point is the separation that they see between the aligner and the teeth and to ensure that the separation is less than or greater than one millimeter, depending on the issue. So that's exactly what Invisalign teaches us. That's more substantial evidence to support the board's conclusion that it would have been obvious if it's not actually shown in the reference.

[00:15:13] Speaker 03:
So I'll go on to the contrast issue. He was saying that the references teach away because allegedly you can't have or Sala doesn't want high contrast or Sala needs high contrast images.

[00:15:28] Speaker 03:
They didn't have any evidence that that was actually correct. On the other side of the ledger, we did have evidence that Sala could use aligners. We actually used the contour detection technique that Sala specifically identified, the canny technique, and our expert was able to say, oh, you know, actually we can detect the clear aligner on teeth. And I would point out to the court that when Sala refers to low contrast, it uses the word blurr. It says, for example, a blur.

[00:15:58] Speaker 03:
It's not even necessarily talking about light, dark. The idea that a clear aligner on teeth can't provide enough contrast for something to detect is demonstrably wrong. It's demonstrably wrong based on the testimony of Dr. Farouk, which the board was entitled to credit. And it's also demonstrably wrong based on the images in Invisalign that we can all see with our own eyes. We can see the distinction between the aligner and the teeth. That's the whole purpose of this endeavor is for someone to look at that. And if the camera can pick up what our eyes can pick up, the camera can pick up the contour difference just as we can pick up the contour difference.

[00:16:33] Speaker 03:
And the last thing I'll say about the contrast issue, and the board pointed this out, there's nothing in Kim that discusses taking an image of a clear liner by itself.

[00:16:45] Speaker 03:
Now, they latch on to that statement and they say, aha, we've been made to prove a negative. That's not true. The board was just trying to analyze what Kim was actually teaching. Their argument was, oh, no, Kim's talking about just a picture of a clear liner on itself. The board was within its rights to say, actually, Kim doesn't say that anywhere. So Kim at least suggests it. Invisalign shows it. We all know why it's needed. It's all needed for clinical evaluation. And Invisalign actually tells us the kind of dimensions we're looking for. And Salah tells us that contour detection can work.

[00:17:16] Speaker 03:
It tells us how to do contour detection. And then Dr. Farouk actually uses a technique in Salah to detect clear aligners and the gap against teeth.

[00:17:30] Speaker 03:
The last thing he discussed, I believe, is the pixel-wise image issue, the pixel-wise data. That's actually an issue for claim seven of the 409 patent, which that limitation is in. I don't think it's even applicable here. I'm not sure what the point he was making, except that he was trying to say that because Meninas uses pixel-wide information, it somehow can't detect contours of teeth. I mean, Meninas is directed to the idea of detecting contours. I think it shows sheep. It might show motorcycles.

[00:18:01] Speaker 03:
It might show something else. It's a contour detection method for all types of things. There's no reason to think it can't work on teeth. There's every reason to think it can. That's what our declarant said. And the other thing is that when Meninas talks about this pixel-wise data, it's the data for every single pixel in the test image, every single one it says. So the images that it tests on has pixel-wide data for every pixel in the image. So if there's some argument that referring to pixel-wise data doesn't tell you about all the contours in the image, that's demonstrably wrong too.

[00:18:40] Speaker 03:
And those are the only three issues that he spoke about was the aligner and the service position, the contrast issue, and the pixel-wise issue, which frankly is not even applicable to this appeal. I'm happy to talk about any other of the issues of which there are many in this case, but if the court has no questions –

[00:18:58] Speaker 00:
Thank you, Mr. Kelly.

[00:18:59] Speaker 03:
Welcome. Thank you, Your Honor.

[00:19:00] Speaker 00:
As you know, one doesn't lose points by not using F. Olin's time. Thank you.

[00:19:06] Speaker 00:
Mr. Sandinato has two minutes for rebuttal. Thank you, Your Honor.

[00:19:13] Speaker 02:
I'll start with the Kim paragraph 66, which, as Your Honor, Judge Toronto pointed out, talks about a clear aligner currently being worn. The point here is that this is about aligner information only. This is not about information only. relating to a liner on teeth. And the fact that the board, by the way, Your Honor, talked about this paragraph and paragraph 83, I believe, which I referred to earlier, which talks about an aligner to be worn or an aligner to be currently worn.

[00:19:48] Speaker 02:
But both of these passages, indisputably, we believe, are about aligner information only and not about information relating to an aligner on teeth or an image of an aligner on teeth.

[00:20:03] Speaker 02:
In terms of the contrast issue, I didn't say that imaging of aligners on teeth is not possible. Clearly, that's something that is possible. The point is, that's not what Salah does. Salah doesn't relate to aligners at all. There's nothing in Salah that even mentions the term aligners. Salah does talk about a harder and prioritizing high contrast areas over low contrast areas, and that was the point that I was making.

[00:20:33] Speaker 02:
On Meninas, it's not about whether the claims require pixel-based determination or don't require pixel-based determination. It's about the claims requiring a deep learning algorithm or a deep learning device that can determine a gap between teeth, and aligners. The claims indisputably do, and a deep learning algorithm that cannot distinguish between objects, and Meninas cannot do that.

[00:21:03] Speaker 02:
Meninas is not class-based, and a deep learning algorithm that can't do that could not distinguish between aligners and teeth, which is necessary to determine a gap. I believe I'm out of time.

[00:21:16] Speaker 00:
Thank you to both counsel. You've given us a lot to chew on.

[00:21:22] Speaker 00:
Admit it.

[00:21:23] Speaker 02:
Thank you, Your Honor.