[00:00:00] Speaker 03:
Our final case this afternoon is number 23-2244, United Services Automobile Association versus People's Bank.

[00:00:10] Speaker 03:
Mr. Rowles.

[00:00:12] Speaker 02:
May it please the court, Anthony Rolls for USAA.

[00:00:16] Speaker 02:
This appeal as to the 136 patent presents two discrete issues.

[00:00:20] Speaker 02:
The first is the board failed to address the dance reference, which was record evidence of prior art introduced in the patent owner response that taught away from the asserted combination of Garcia's mobile phone camera with Richardson's OCR of the routing number on the check, which is a requirement of all the claims.

[00:00:38] Speaker 02:
The claim set one requires specifically OCR of the routing number.

[00:00:42] Speaker 02:
Claims 14 through 18 require OCR of deposit information which the board construed to include the routing number as well as other information.

[00:00:51] Speaker 02:
Garcia admittedly doesn't teach determining this routing number or other information via OCR.

[00:00:57] Speaker 02:
They relied on Richardson for that and the dance reference

[00:01:01] Speaker 02:
specifically teaches that relying on optical character recognition to obtain information from a document captured by a handheld camera, as opposed to a scanner or a mounted device, won't work.

[00:01:13] Speaker 02:
The image will appear acceptable to the user.

[00:01:16] Speaker 02:
The user will perceive it as being of adequate resolution and looking acceptable.

[00:01:21] Speaker 03:
What's the date of the dance reference?

[00:01:23] Speaker 02:
2000.

[00:01:23] Speaker 02:
And the date of Garcia?

[00:01:26] Speaker 02:
Garcia is, I believe, 2004 or 2005.

[00:01:30] Speaker 02:
So Dance is prior to Garcia, Your Honor.

[00:01:34] Speaker 03:
So technology changes.

[00:01:37] Speaker 02:
It does.

[00:01:38] Speaker 02:
And perhaps the board could have relied on that in assessing Dance.

[00:01:41] Speaker 02:
It did not.

[00:01:42] Speaker 02:
There was no finding in the final written decision about the evolution of technology, about the applicability of Dance's teaching away to Garcia at the time of the invention.

[00:01:52] Speaker 02:
The final written decision is simply absent of any discussion of dance of the teaching of dance and its applicability to the combination asserted here and so that on its own.

[00:02:04] Speaker 02:
is an error of the board.

[00:02:07] Speaker 02:
When the patent owner makes of record prior art that teaches away from the asserted combination in the patent owner response, the board must address it.

[00:02:16] Speaker 02:
It does not need necessarily to accept it as excluding obviousness, but it must address it and explain why the dance reference would not negate the motivation to combine the references presented in the petition.

[00:02:30] Speaker 02:
It failed to do so here.

[00:02:31] Speaker 02:
And the second error made by the board relates just to claims 14 through 18 of this patent.

[00:02:37] Speaker 02:
And those are the claims that have the signature identification procedure limitation.

[00:02:41] Speaker 02:
And so the claims as construed by the board, and no one disputes the claim construction,

[00:02:47] Speaker 02:
limit the invention to performing a specific process on the back of the check.

[00:02:52] Speaker 02:
It has to detect a mark in a particular location on the back of the check, the endorsement location, and without performing a quote unquote signature identification procedure, which has been construed to include identifying the presence of a signature on the back of the check.

[00:03:11] Speaker 02:
And so the claim, when you put all that together, the claims describe this process where rather than looking for a signature on the back of the check, you look for the presence of a mark.

[00:03:21] Speaker 03:
Even Randall doesn't look just for a signature, it looks for an endorsement, which could be a signature or some other mark, right?

[00:03:27] Speaker 02:
Randall teaches looking for an endorsement.

[00:03:29] Speaker 02:
As the petition says, an endorsement search is what's taught by Randall and is going to be combined into the Garcia reference.

[00:03:38] Speaker 02:
Now, it's indisputed that in the vast majority of cases, an endorsement was understood to be a signature, but that in certain cases, in commercial checks, a stamp could be used in lieu of a signature.

[00:03:50] Speaker 02:
So I agree with your honor on that.

[00:03:52] Speaker 02:
But if we look at Randall's teaching to search for the endorsement on the back of the check, it's teaching a process to look for something specific, to look for an endorsement, which is a signature or potentially something used in lieu of a signature, a stamp.

[00:04:05] Speaker 02:
That's different from what's recited in the claim, which is to look simply for a mark in a particular area on the backside of the check.

[00:04:13] Speaker 03:
Looking for a stamp isn't looking for a signature.

[00:04:16] Speaker 02:
I agree that a stamp is not a signature.

[00:04:19] Speaker 02:
That's true.

[00:04:20] Speaker 02:
What Randall does not teach is just looking for a stamp.

[00:04:23] Speaker 02:
There's no suggestion of that.

[00:04:24] Speaker 02:
There's no argument made by petitioner that Randall teaches alternatives.

[00:04:28] Speaker 02:
You could look for a signature or a stamp.

[00:04:30] Speaker 02:
The admission by PNC in the red brief at 39 is that in the vast majority... You mean that you read the claim as saying it's looking for a mark, it's not a signature?

[00:04:41] Speaker 02:
I think looking for a mark is something different from looking for a signature or looking for a mark.

[00:04:47] Speaker 03:
But a signature can be a mark.

[00:04:49] Speaker 02:
I think to be precise, one makes a signature by making marks or a series of marks on the back of a check.

[00:04:55] Speaker 02:
I don't think I would agree that a signature is a mark.

[00:04:58] Speaker 02:
Maybe it's comprised of marks.

[00:05:00] Speaker 02:
But the process claimed in the patent is rather than looking for something specific like a signature, the mere presence of a mark, it could be a stray mark, it could be an X or a stray pencil mark if it's in the endorsement location.

[00:05:17] Speaker 02:
would be enough to pass that step of processing in the claimed invention.

[00:05:22] Speaker 02:
That would not be the case for Randall.

[00:05:24] Speaker 02:
Randall would not find an endorsement, and so Randall would reject the check.

[00:05:28] Speaker 02:
Conversely, you could have situations where the patented invention differs from Randall.

[00:05:33] Speaker 02:
There could be a signature that's not in the endorsement location.

[00:05:36] Speaker 02:
The 1 through 6 patent would not consider that as passing its processing step.

[00:05:41] Speaker 02:
Randall would, because it would find the endorsement.

[00:05:43] Speaker 02:
on the back of the check.

[00:05:44] Speaker 02:
These are simply two different processes, and the patent identifies this different process as being advantageous.

[00:05:52] Speaker 02:
This is at appendix 96 in the 136 patent, column 10, lines 2 through 6, when it explains this alternative embodiment of looking for, in the specification it says looking for just a mark or signature.

[00:06:06] Speaker 02:
what ends up getting claimed is just a mark, in the endorsement location on the back of the check without conducting any further signature identification procedures, that avoids the problem of trying to look for something like a signature that's sometimes illegible.

[00:06:22] Speaker 02:
You can't tell what it is.

[00:06:23] Speaker 02:
You can't tell that it's a signature.

[00:06:25] Speaker 02:
Using the claimed process, you avoid that problem.

[00:06:28] Speaker 02:
You're looking for just the presence of a mark at the cost, perhaps, of false positives where you detect a stray

[00:06:34] Speaker 02:
a pen stroke as an endorsement, but that's considered something that's advantageous of the invention, and that's simply not taught by Randall.

[00:06:43] Speaker 02:
And importantly, there's not an argument in the petition that a person of skill would have modified Randall to say, well, let's just look for stamps, or let's just look for a certain type of stroke and not a signature.

[00:06:57] Speaker 02:
That's not the argument the petitioner made.

[00:07:00] Speaker 02:
The petition's theory was that Randall taught two different things you could do.

[00:07:08] Speaker 02:
You could look for a signature on the front side of the check and you could also optionally compare it to one that's on file.

[00:07:14] Speaker 02:
And the petition reasoned that

[00:07:16] Speaker 02:
The same would apply to the backside of the check.

[00:07:18] Speaker 02:
You could look for the endorsement, but not necessarily compare the signature to that on file.

[00:07:23] Speaker 02:
And that would be advantageous because you avoid this complex process of looking up the signature that's on file.

[00:07:28] Speaker 02:
That was the theory presented in the petition.

[00:07:29] Speaker 02:
The claim, as construed, doesn't require that.

[00:07:33] Speaker 02:
The signature identification procedure that's excluded from the claim is broader than looking up a signature on file.

[00:07:40] Speaker 02:
It includes even looking for the presence of a signature at all.

[00:07:44] Speaker 02:
And so that results in this flaw in the petition, which the board

[00:07:49] Speaker 02:
ignored in its decision.

[00:07:51] Speaker 02:
What the board did was say, well, Randall doesn't specifically say, look for a signature, and that's enough to satisfy the claim limitation.

[00:08:00] Speaker 02:
But Randall is teaching looking for an endorsement known to be a signature or potentially a stamp in lieu of a signature.

[00:08:08] Speaker 02:
So it's at least going to be looking for the presence of a signature on the back of a check.

[00:08:13] Speaker 02:
And that puts it outside of the scope of the 136-pack claims.

[00:08:17] Speaker 02:
As to the teaching away point, I think the problem here is that

[00:08:25] Speaker 02:
There's a lot of evidence in the record about, and we've heard about it today, evidence in the record about difficulties with using a camera phone and an image captured by a camera phone for deposit of a check.

[00:08:37] Speaker 02:
And in this case, the combination asserted is slightly different from some of the other cases.

[00:08:42] Speaker 02:
They've combined Garcia with this patent Richardson.

[00:08:45] Speaker 02:
And Richardson is describing a commercial bank processing system where high-speed scanners are taking images of checks and it's reading information from the checks and processing it.

[00:08:55] Speaker 02:
And so when you can now petition argues that nonetheless a person of skill would have thought you could apply that technology to Garcia's images and they're within the right to argue that but when other evidence of record of the prior art says

[00:09:09] Speaker 02:
While those techniques work for scanners, which is what Dan says, while those techniques work for flatbed scanners, when you attempt to do it with a handheld camera without using other techniques described in Dan's, which aren't in the combination, then what happens is the user will believe the image is adequate, in this case submit it for deposit,

[00:09:29] Speaker 02:
only to learn later that the processing cannot be performed.

[00:09:33] Speaker 02:
And then, Dance explains, the user will be unlikely to ever use such a system again because they sent an image they believed to be adequate for processing and only learned later that it was inadequate for processing.

[00:09:44] Speaker 02:
These are teachings that would specifically discourage a person of skill from making this combination of the camera deposit system with Richardson's scanner-based OCR of the critical check information.

[00:09:57] Speaker 02:
And the board, you can search the entire final written decision.

[00:10:01] Speaker 02:
You won't find any mention of those teachings.

[00:10:03] Speaker 02:
The closest the board comes to addressing this issue is it refers to an argument made by USA that a prior board decision with a petitioner called MyTech involved the dance reference.

[00:10:19] Speaker 02:
And the board says, well, we put no weight on the MyTech decision because it was a different references, different art.

[00:10:26] Speaker 02:
Yes, Dr. Cunningham.

[00:10:27] Speaker 01:
A couple things.

[00:10:28] Speaker 01:
One, can you just confirm that you're not covering up the mindset that is being to me?

[00:10:36] Speaker 01:
I'm getting a lot of feedback.

[00:10:37] Speaker 01:
That's one of the minor things that the paper said.

[00:10:39] Speaker 01:
And I totally get it.

[00:10:40] Speaker 01:
I've been in your shoes arguing and how... My apologies.

[00:10:43] Speaker 01:
I'll hear you clearly.

[00:10:45] Speaker 01:
And then number two, what is the distinction, if any, that you're drawing between the mindset argument and the landscape argument?

[00:10:53] Speaker 02:
In this case, the fact that the MyTech panel came out another way on an institution decision, we placed no weight on that.

[00:11:02] Speaker 02:
We're not arguing that that's a basis to find error in this panel decision.

[00:11:07] Speaker 02:
We did argue that in the preliminary response, unsuccessfully.

[00:11:12] Speaker 02:
When we got to the patent response, what we said is, well, let's look at the dance reference on its own.

[00:11:17] Speaker 02:
It's prior art.

[00:11:18] Speaker 02:
It's of the time, within a few years.

[00:11:20] Speaker 02:
And the teachings of that reference directly contradict the stated motivation to combine in this case.

[00:11:26] Speaker 02:
And that's a reason why the prior art teaches away from the combination.

[00:11:31] Speaker 02:
When we got to the final written decision, the board didn't address that argument.

[00:11:36] Speaker 02:
They just repeated what they said in the institution decision that MyTech is different.

[00:11:39] Speaker 02:
The panel was addressing a different combination.

[00:11:42] Speaker 02:
The Singfield reference was an issue, and in the MyTech reference, there were a lot of differences.

[00:11:46] Speaker 02:
But that's immaterial to the argument that we're making here.

[00:11:50] Speaker 02:
The prior art of record is dance.

[00:11:53] Speaker 02:
It's material to the combination.

[00:11:54] Speaker 02:
It refutes the motivation to combine.

[00:11:57] Speaker 02:
It needed to be addressed by the board here on its own merits, regardless of what the prior panel in an institution decision center didn't say.

[00:12:06] Speaker 02:
That being said, the findings of the prior panel were that Dance taught away from using a camera to take images of a check and deposit it.

[00:12:16] Speaker 02:
So we're consistent with what the board found before, but the court need not go there on this appeal.

[00:12:24] Speaker 02:
The dance reference on its own is record evidence in support of patent and response and patenters theory unaddressed by the board.

[00:12:32] Speaker 02:
And that is something that the board is not within its rights to do.

[00:12:36] Speaker 02:
It cannot simply ignore

[00:12:37] Speaker 02:
a substantive argument made in the patented response without explaining why it finds it not persuasive, why it finds it waived, some reason why.

[00:12:45] Speaker 03:
It has to address every single argument?

[00:12:48] Speaker 02:
I think there's a threshold, and this court's cases indicate that.

[00:12:51] Speaker 02:
But this is, again, this is not a footnote where we said, you know, CEG dance in the MyTech IPR.

[00:12:57] Speaker 02:
This is four pages of the patented response going through the teachings of dance and explaining why they resulted in, quote, no motivation to combine.

[00:13:04] Speaker 02:
Garcia with Richardson.

[00:13:07] Speaker 02:
And this is on the other side of that line where it's a substantive argument that the board needed to address to explain why dance didn't contradict the motivation to combine.

[00:13:18] Speaker 02:
Not that literally everything in a patented response needs merits of response, but I think this was a significant argument in the patented response that required such an adjudication.

[00:13:30] Speaker 02:
I see I'm into my rebuttal time.

[00:13:32] Speaker 03:
I'll give you two minutes.

[00:13:33] Speaker 03:
Thank you, Your Honor.

[00:13:35] Speaker 03:
Mr. Danford.

[00:13:49] Speaker 04:
Good afternoon again.

[00:13:50] Speaker 04:
Andrew Danford for PNC.

[00:13:52] Speaker 04:
I'll start with the issue with respect to the dance reference and direct the

[00:13:59] Speaker 04:
court to appendix page 40 where the board discussed the MyTech

[00:14:06] Speaker 04:
proceedings and the art and arguments that were raised in those proceedings and distinguished it from the art and arguments that are at issue here because this proceeding involved the disclosures in Garcia, which has the explicit disclosure of check images captured by a mobile phone that are then processed by OCR.

[00:14:27] Speaker 04:
The board when it was talking about MyTech was describing the dance reference in exactly the same way.

[00:14:33] Speaker 04:
that the patent owner's response did.

[00:14:35] Speaker 04:
And I would direct you to pages 596 to 598 of the appendix where USAA addresses the dance reference specifically citing to MyTech.

[00:14:49] Speaker 04:
And so it's clear when you read the patent owner's response against the final written decision that what the board is talking about at page 40 of its final written decision is addressing

[00:15:00] Speaker 04:
those arguments with respect to dance and explaining why they don't change the outcome here.

[00:15:06] Speaker 04:
And I'd also make the point too, and I think this is why, part of the reason why the board

[00:15:12] Speaker 04:
stated that the dance reference and the art at issue in the MyTech proceeding don't change the outcome here is because we don't need a motivation to combine to get a check image captured by a mobile phone that is then processed using OCR because that is the explicit disclosure of Garcia that is all in one reference.

[00:15:32] Speaker 04:
That is what Garcia teaches.

[00:15:34] Speaker 04:
So those issues that were at issue in the MyTech proceeding, the issues that are addressed by the arguments with respect to dance,

[00:15:41] Speaker 04:
just don't have an application here because the art is different, the arguments are different, and there's not this whole issue about a motivation to combine to get OCR of those images just isn't in play given the state of the art in 2003 with Garcia.

[00:15:59] Speaker 04:
I would also, to the extent there is an argument about the motivation to combine with Richardson, we think that it's readily supplied by the evidence that the board cited that was presented in the petition and supported by our expert Dr. Noble, in particular

[00:16:15] Speaker 04:
What we have here is Garcia, which teaches the check images captured by mobile phone that are then processed by OCR.

[00:16:22] Speaker 04:
And the only question at that point is, well, what information do you actually read off the check using OCR?

[00:16:28] Speaker 04:
The board concluded the obvious thing to do is to read the numbers printed at the bottom of the check, which are any special font that enables them to be machine read.

[00:16:37] Speaker 04:
and that is critical information for processing the check deposit.

[00:16:41] Speaker 04:
This is an issue that we've seen crop up across a number of these appeals today.

[00:16:45] Speaker 04:
The board made findings about why a person of skill in the art would be motivated to apply OCR to that given the importance of that information and the risk of error if you were to actually have that information entered manually because this is ultimately just a string of numbers.

[00:17:06] Speaker 04:
Turning to the signature issue, the signature limitation is disclosed and taught by Randall, which teaches handling the endorsement on the back of the check differently from how you do the signature on the front.

[00:17:23] Speaker 04:
So all that Randall teaches, if you look at appendix page 1224, the figure on that page has box 124 that says, endorsement, is it there?

[00:17:36] Speaker 04:
That is just a yes-no check to determine whether there is an endorsement mark present in the endorsement space.

[00:17:42] Speaker 04:
It is not checking for a signature, and checking for the presence of an endorsement is different from checking for the presence of a signature, and it's different from checking whose signature it is.

[00:17:53] Speaker 04:
And this wasn't an instance of the board relying on the references near silence on this issue.

[00:17:58] Speaker 04:
The board recognized that the disclosure in Randall

[00:18:04] Speaker 04:
for how you handle the endorsement on the back of the check is different from how it discloses how you handle the signature on the front of the check, where there is an additional step where the signature is actually compared against the signature on file, which is exactly the kind of signature verification step that the claim in the 136 patent says not to do.

[00:18:23] Speaker 04:
Our expert, Dr. Noble, said a person of skill in the art would understand that what the Randall reference is disclosing with respect to the back of the check and the endorsement

[00:18:34] Speaker 04:
is exactly what the 136 patent claims in view of those differences and how the reference handles the front and back of the check.

[00:18:41] Speaker 04:
And the board was well within its rights to accept that as reasonable and explained and supported by the record.

[00:18:48] Speaker 04:
So unless the court has any further questions for me, we'd submit the rest of this on the papers and just ask that you affirm.

[00:19:09] Speaker 02:
Thank you, Your Honor.

[00:19:11] Speaker 02:
I'll start with the dance reference, and I would direct the court to Appendix 34, where... Excuse me for a moment.

[00:19:27] Speaker 02:
At Appendix 34, the first full paragraph beginning petitioner notes, there was an argument that this dance reference is not an issue here because Garcia itself teaches OCR of the information required.

[00:19:41] Speaker 02:
Remember that the claim requires specifically OCR of the routing number at minimum.

[00:19:45] Speaker 02:
The board here says petitioner notes that Garcia's disclosure of check imaging and the incorporation of OCR equipment

[00:19:51] Speaker 02:
does not explicitly state the contents of the check includes the routing number, but that Richardson discloses OCR of the microline and then relies on Richardson to supply it.

[00:20:01] Speaker 01:
Yes, Judge Cunningham.

[00:20:12] Speaker 02:
Well, there's certainly no reference to dance.

[00:20:15] Speaker 02:
To the extent it addresses my tech, it mentions that the my tech decision involved a different prior combination with a different motive to combine.

[00:20:25] Speaker 02:
I don't dispute that, but the motivation to combine here of Garcia with Richardson is impacted by the teachings of the dance reference, and the board does not grapple with that.

[00:20:37] Speaker 02:
It apparently cast that aside based on the MyTech IPR involving different issues.

[00:20:43] Speaker 03:
It involved dance, right?

[00:20:44] Speaker 02:
It involved dance among other references and a different base reference and so different issues.

[00:20:52] Speaker 02:
But nonetheless, the question here is, and the board there found the dance taught away from using a check image captured by a digital camera.

[00:21:01] Speaker 02:
setting that aside, here the question is does the combination and the motivation relied on what I just read from you from appendix 34, they're relying on the combination of Garcia with Richardson, does DANCE discredit that motivation to combine those two references?

[00:21:17] Speaker 02:
And it does, or at least the board failed to explain why it did not.

[00:21:21] Speaker 02:
And a passing reference to the MyTech IPR being different doesn't

[00:21:24] Speaker 02:
satisfy the board's burden under the APA to resolve that.

[00:21:30] Speaker 00:
Can I just ask you one technical question?

[00:21:32] Speaker 00:
Because they're not citing to the Mitac case.

[00:21:35] Speaker 00:
They are explicitly citing to the discussion of IPR proceedings pursued by a third-party Mitac.

[00:21:43] Speaker 00:
And they go on to cite Patnona response 55 to 56.

[00:21:47] Speaker 00:
Is that the portion in which you were discussing dance?

[00:21:50] Speaker 02:
The portion, the teachings of dance I discussed earlier are actually on Patnona response 57 to 58.

[00:21:57] Speaker 02:
At 55 to 56, we introduced the MyTech IPR, explained that dance was at issue there, and acknowledged that the MyTech IPR involved different combinations and different motivations to combine.

[00:22:10] Speaker 02:
So I believe that's the

[00:22:11] Speaker 02:
And that's what petitioner said in its reply at 16.

[00:22:14] Speaker 02:
Okay.

[00:22:14] Speaker 00:
And then they later cite Pat Noner response 57 and 58.

[00:22:22] Speaker 02:
Pat Noner response 57 to 58, we discussed dance as well as Richardson teaching away.

[00:22:27] Speaker 02:
And so the board addresses the allegation that Richardson taught away.

[00:22:33] Speaker 02:
I'm sorry, it's not here, but 57 to 58 addresses dance in addition to other references, including Richardson.

[00:22:40] Speaker 02:
answer your question.

[00:22:43] Speaker 02:
On the signature point, there was an argument made that the only difference in Randall from the front side of the check to the back is that on the back, it's not comparing anything to a signature on file.

[00:22:55] Speaker 02:
And just wanted to remind the court that the agreed claim construction of signature identification procedure, which is what is the negative limitation, includes both comparing a signature to something on file and determining the presence of the signature.

[00:23:09] Speaker 02:
So it's OK.

[00:23:10] Speaker 03:
I think we're out of time.

[00:23:11] Speaker 03:
Thank you.