`
`
`
`
`
`
`UNITED STATES PATENT AND TRADEMARK OFFICE
`
`
`
`BEFORE THE PATENT TRIAL AND APPEAL BOARD
`
`___________
`
`ELECTION SYSTEMS & SOFTWARE, LLC,
`
`Petitioner,
`
`v.
`
`HART INTERCIVIC, INC.,
`
`Patent Owner.
`___________
`
`PGR2020-00031
`Patent 10,445,966 B1
`___________
`
`Record of Oral Hearing
`Held: July 6, 2021
`_____________
`
`Trials@uspto.gov
`571-272-7822
`
`
`
`
`
`
`
`
`
`Before CHARLES J. BOUDREAU, AMANDA F.WIEKER, and
`JASON M. REPKO, Administrative Patent Judges.
`
`
`
`
`
`
`
`
`
`
`
`
`
`PGR2020-00031
`Patent 10,445,966 B1
`
`
`APPEARANCES:
`
`ON BEHALF OF THE PETITIONER:
`
`
`ROBERT EVANS, ESQUIRE
`MICHAEL HARTLEY, ESQUIRE
`Lewis Rice, LLC
`600 Washington Avenue
`Suite 2500
`St. Louis, MO 63101
`
`
`
`ON BEHALF OF THE PATENT OWNER:
`
`
`BRIAN OAKS, ESQUIRE
`MARK SPEEGLE, ESQUIRE
`Baker Botts, LLP
`2001 Ross Avenue
`Dallas, TX 75201
`
`
`The above-entitled matter came on for hearing on Tuesday, July 6, 2021,
`commencing at 2:00 p.m., EDT, at the U.S. Patent and Trademark Office, by
`video/by telephone.
`
`
`
`
`
`
`
`
` 2
`
`
`
`PGR2020-00031
`Patent 10,445,966 B1
`
`
`
`
`
`
`
`
`
`P R O C E E D I N G S
` - - - - -
`JUDGE REPKO: This is an oral hearing for PGR 2020-
`
`00031. I'm Judge Repko. I'm joined by Judges Boudreau and
`Wieker. First of all, thank you for your flexibility and
`understanding in having to present your arguments at a video
`hearing. If any technical difficulties arise, please email or call
`the hearing staff. Our primary concern is your right to be heard
`so if you have any technical problems you feel undermines your
`ability to represent your client in any way, please let us know
`immediately and as soon as we're aware that someone gets
`disconnected we'll pause the hearing while they reconnect. Also
`when you are not speaking, please mute yourself. And, when
`you do speak, please identify yourself at the start of your
`remarks so the court reporter can get an accurate transcript.
`
`At the conclusion please remain on the line so the court
`reporter -- you can answer any of the court reporter's questions.
`Feel free to present yourself however you feel comfortable, that
`means it's okay to sit or stand. There are members of the public
`listening to the oral hearing today. So if there's any confidential
`information, I don't think there is, but if there is, you need to let
`us know so we can make sure that we don't violate that
`confidentiality. At this time, we'd like counsel to introduce
`themselves and anyone with them, and I will begin with
`Petitioner's counsel.
`
`
`
`
` 3
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`
`
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
`PGR2020-00031
`Patent 10,445,966 B1
`
`MR. EVANS: Thank you, Your Honor. My name is Robert
`
`Evans. I'm here with Michael Hartley. We're both with the firm
`Lewis Rice. We're appearing on behalf of Petitioner.
`
`JUDGE REPKO: Will you be presenting the argument fully
`today or handing it off?
`
`MR. EVANS: I will be presenting the Section 101
`argument for the first ten or fifteen minutes and then I will hand
`it to Mr. Hartley who will handle the 102 and 103 arguments.
`
`JUDGE REPKO: Thank you.
`
`MR. EVANS: And 112.
`
`JUDGE REPKO: Okay. And Patent Owner's counsel.
`
`MR. OAKS: Yes, Your Honor. My name is Brian Oaks. I
`represent Patent Owner Hart Intercivic and here with me today is
`co-counsel Mark Speegle. To answer your question of the
`Petitioner I will be discussing the grounds 1 and 2 of the
`Petitioner's petition and Mr. Speegle will be addressing the
`remaining grounds.
`
`JUDGE REPKO: Okay. Thank you. So our Hearing Order
`gave each party one hour to present their arguments. Since we
`don't have a clock for everyone to look at, I'll try to time you and
`give you a warning when you have about five minutes remaining.
`Petitioner's counsel will begin followed by Patent Owner's. Both
`parties may reserve some rebuttal time, but a party may not
`reserve more than half their total time unless there's some special
`circumstances. If you have some objections please raise them
`
`
`
`
` 4
`
`
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
`PGR2020-00031
`Patent 10,445,966 B1
`
`during your rebuttal. We do note that Patent Owner's objected to
`Petitioner's demonstratives. At this time we're going to reserve
`ruling on that objection. We will allow presentation of the
`arguments and demonstratives here today, but ultimately, we will
`not consider them in our final written decision if we determine
`that it would be improper. Generally demonstratives are not
`evidence and will not be relied upon as evidence. It is visual
`aids about the arguments and evidence presented in the papers,
`but they cannot be used to advance new arguments or introduce
`evidence not previously presented in the record.
`
`So with that, I'm going to invite Petitioner's counsel to
`begin, and I need to know how much time you will reserve for
`your rebuttal.
`
`MR. EVANS: We'd like to reserve 15 minutes for rebuttal,
`Your Honor.
`
`JUDGE REPKO: Okay. Thank you. You may begin.
`
`MR. EVANS: Thank you, and good afternoon. I'm going
`to address Section 101 and then Mr. Hartley will address
`Sections 102, 103 and 112. I'd like to start by noting that all the
`claims in this patent are directed to the abstract idea of voting
`and verifying the addressee and authenticity of the printed record
`of the vote. If we look at slide 4, please. In the case Voter
`Verified, Inc. v. Election Systems & Software, LLP, the Federal
`Circuit found that,
`
`"Humans have performed this fundamental activity of
`
`
`
`
` 5
`
`
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
`PGR2020-00031
`Patent 10,445,966 B1
`
`voting that forms the basis of our democracy for hundreds of
`years."
`
`Accordingly, claims that as a whole are drawn to the
`concept of voting, verifying the vote and submitting the vote for
`tabulation are ineligible. So just keeping that framework in mind
`as we go forward would be helpful.
`
`If we can turn please to slide 5. What we have there are
`independent claims 1, 6 and 11 and they're highlighted in yellow
`to show that they have a common set of introductory claim
`elements recording the voter's vote selections, generating a
`printed vote record of those selections, scanning the printed vote
`record, utilizing optical character recognition to create a data set
`of those vote selections and then generating a cast vote record
`from the data set. But in summary those are the common
`elements of all the claims in this patent, the independent claims,
`and then we see in claim 1 is that they use a comparison with an
`election dictionary to ensure the accuracy of the optical character
`recognition.
`
`In claim 6 they use a barcode with the voter's selections in
`the bar code to ensure the accuracy of the vote and in claim 11
`they use a hash function where they hash the original data and
`then they hash the data on the printed record and they put the
`hash of the data on to the face of the ballot so you can compare
`the hash as originally presented with the original vote selection
`against the hash of the later reviewed document and if the hashes
`
`
`
`
` 6
`
`
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
`PGR2020-00031
`Patent 10,445,966 B1
`
`match then you know there has been no changing of the ballot.
`That's standard digital signature technology that's been around to
`self-authenticate documents for a long time.
`
`So we submit that the recitation of the abstract idea that
`was given is an accurate one. We note that in the Patent Owner's
`papers they took issue and said we should have used the word
`integrity instead of authenticity. I submit that in the context of
`the patent and in the context of the claims, that those words are
`synonyms here. Secondly, they're complaining that we didn't
`have all the claim elements in our recitation of the abstract idea
`and that's not the purpose of the abstract idea. The purpose of
`the abstract idea is to be higher level statement of what's the idea
`for the claim, what's the idea for the patent and I think we
`captured that here (indiscernible.)
`
`JUDGE REPKO: So if the -- excuse me counsel. So if the
`abstract idea is voting, are there specific limitations that are
`directed to the act of voting? I know you're saying the abstract
`idea is maybe a high level generalization of the claim, but can
`you point to specific limitations that capture that abstract idea?
`
`MR. EVANS: Yes. Well it's not just voting, it's also that
`verifying the accuracy and authenticity of the printed record of
`the vote. So it's voting but requiring the further elements that
`you have to verify that the printed record you made -- so you
`have to make a printed record and then you have to verify the
`accuracy and authenticity of that printed record and that's what
`
`
`
`
` 7
`
`
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
`PGR2020-00031
`Patent 10,445,966 B1
`
`covers the remaining claim elements I believe.
`
`JUDGE REPKO: Well, is verifying the vote is performed
`by a machine, it's a machine's responsibility, it's not a person's;
`right? Is that the case here in these claims too? Or do you view
`it as broad enough to encompass a person verifying the vote?
`
`MR. EVANS: No. The verification here takes place when
`the machine uses a election dictionary with the OCR or the
`machine uses a barcode to compare the contents of the barcode
`against the scanned election selections that the voter may or will
`use a hash function to verify the accuracy of the ballot and so
`those are all machine functions performed by a generic computer.
`
`JUDGE REPKO: They're a part of the abstract idea, is that
`what you're saying? Is that correct?
`
`MR. EVANS: Yes.
`
`JUDGE REPKO: Okay. So hashing and OCR, that's --
`you're saying in your view is abstract?
`
`MR. EVANS: Yes. It's using a generic computer to
`perform generic functions that have been used in the computer
`space forever and have even been used in the voting space in
`many places. So that gets to sort of the second question which
`is, is this a practical application and in the context of practical
`applications for computers, you have to ask yourself did they
`improve the function of the computer or did they not and if they
`didn't, well that simply using a generic computer is not a
`practical application.
`
`
`
`
` 8
`
`
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
`PGR2020-00031
`Patent 10,445,966 B1
`
`Here what we find is that OCR has been used for decades.
`
`Shamos's declaration is unrebutted in paragraphs 56 and 57.
`Exhibit 1001 showed OCR in fact in 1967. Levenshtein, the
`technique you use to determine whether or not the OCR is
`accurate has been around since the '60s. The use of dictionaries,
`Heilpur, McClure, Toledo, Exhibit 1002, Exhibit 1027, and then
`Heilpur -- I don't have the exhibit number here, but Shamos at
`paragraphs 57, 58 and 72 of his first declaration explained how
`those use of election dictionaries in the context of voting and
`OCR in the ballot were well known well before this patent.
`Keller (indiscernible) --
`
`JUDGE REPKO: So you're saying that these are -- you're
`saying these are known functions but you're also saying it's part
`of the abstract idea? Is that your position?
`
`MR. EVANS: Yes. It's my position that these are known,
`conventional functions of generic computer components that do
`not contribute anything other than what was old, anything other
`than just a generic component Lego chips, plug them together
`and then they go.
`
`JUDGE REPKO: Okay. Thank you counsel.
`
`JUDGE BOUDREAU: This is Judge Boudreau. I just
`wanted to clarify that -- so your position is that those are known
`functions and that they could all be performed either in the
`human mind or by a human using pen and paper; is that correct?
`
`MR. EVANS: It's certainly a pen and paper was the
`
`
`
`
` 9
`
`
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
`PGR2020-00031
`Patent 10,445,966 B1
`
`identical thing here, the computer simply replaces the human
`mind, replaces the pen and paper. Absolutely.
`
`JUDGE BOUDREAU: Thank you.
`
`MR. EVANS: Keller -- thank you. Keller, Exhibit 1003,
`shows that in 2005 they were using a printed vote record in
`human readable text using an OCR friendly font. So again you
`have OCR as a printed vote record well before this patent. So
`again it's just a generic computer function and, you know, if you
`saw on the record it's known that when you OCR you're going to
`have errors because print may not be perfectly clear everywhere.
`Some letters get read as something else. They talk about OCR
`friendly fonts but not every font is OCR friendly and so you need
`to have a dictionary and when you have that dictionary you're
`looking for the words you expect to see. Obviously if there are
`no elephants on the ballot you wouldn't expect elephant to be in
`your dictionary. If George Washington is on the ballot, you'd be
`looking for George Washington, did someone vote for George
`Washington and so this idea of an election dictionary while it
`sounds specific, dictionaries have been used since the beginning
`of OCR-ing because that's how you make sure that your OCR
`doesn't just produce a jumble of letters and what you do is if the
`letters match to a word in the dictionary, you accept the word as
`Dr. Shamos explained and if the letters are close, you take the
`closest combination of letters to the closest word in the
`dictionary. That's the Levenshtein algorithm that's been used
`
`
`
`
` 10
`
`
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
`PGR2020-00031
`Patent 10,445,966 B1
`
`since the '60s.
`
`JUDGE REPKO: Is that a --
`
`MR. EVANS: And as Dr. Shamos explained --
`
`JUDGE REPKO: -- excuse me, is that -- that part of the
`abstract idea? Is that trying to solve perhaps a technical
`problem? Is what you're saying with respect to fonts and, you
`know, transforming something in the physical space to the digital
`world, is that a solution to a technical problem, or is that merely
`part of the abstract idea?
`
`MR. EVANS: It's part of the abstract idea because it's a
`well known generic computer function that was known well
`before this patent and so in asking the question is there a
`practical application here, you know, was this an improvement
`on the computer or was this just using generic computer
`functions? I submit that since all of these technologies well
`preceded this patent they're just generic functions and they don't
`count in terms of practical application under the controlling test.
`
`If we look at barcodes they've been used since at least
`1949. Dr. Shamos gave numerous examples of barcodes in the
`election voting ballot context in paragraphs 81 and 133 of this
`declaration. They even had hash functions. They even used
`ballots in elections. He gave examples of that in paragraphs 87,
`88 and 133 of his declaration. I'd also direct you to Exhibits
`1042, 1043, 1025 and 1008. So all of these functions are generic
`functions that well preceded this patent and so when you just
`
`
`
`
` 11
`
`
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
`PGR2020-00031
`Patent 10,445,966 B1
`
`take old generic functions and you connect them together with
`generic hardware you don't have a practical application and that's
`where you end. If you look at the patent itself it admits at
`column 4, lines 13 to 15 that the scanner it presents is
`conventional. Keller 4, lines 30 to 35, the marking device
`including the hardware, software, memory, processor, input
`devices, those were all "known in the past." That's what the
`patent says. Keller 4, lines 45 to 46, they talk about a thermal
`printer. You know, we've got thermal printers in grade school
`and before and in column 6, lines 15 to 18 they said the barcode
`can be of any type. So this is the patent saying this is all generic
`stuff as well and so that --
`
`JUDGE REPKO: I was going to say we need to consider
`whether something is well understood, routine and conventional
`under our guidelines, which in the first part of the test we're
`looking for an abstract idea, and if we determine that there is,
`you know, some sort of technology here being improved or
`whatever, do we actually even need to reach that issue? And I
`guess my question for you would be, you know, looking at all of
`these things from the background that you're citing, it all seems
`to be talking about improvements, you know, alleged
`improvements to a computer—the way it functions. So does that
`take it outside the realm of an abstract idea at that point when
`you're trying to focus on improving the computer itself
`regardless of whether we have components individually that are
`
`
`
`
` 12
`
`
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
`PGR2020-00031
`Patent 10,445,966 B1
`
`well understood, routine and conventional?
`
`MR. EVANS: I don't believe so because the case law is
`very clear that when you have simply a combination of generic
`computer components performing generic functions measured as
`of the time of the invention that you have to analyze it under
`Section 101 and when you find you yourself in that situation
`you're not improving a computer if the computer has already
`been improved. So if the improvement has already been made by
`others long before you and it's well known and conventional,
`well then there's nothing left for you to improve. The
`improvement has been made. You can't say hey, I just improved
`a computer by taking something old and off-the-shelf.
`
`When we asked their expert, you know, what's up with
`these things, he admitted that at slide 14 that a general purpose
`computer would be the thing you use the ballot marking device.
`He admitted at slide 16 that the barcode could be an open source
`barcode. He admitted further there that the OCR process that the
`patent didn't disclose anything specific. That's because OCR
`processors are available off- the-shelf. He admitted at page 17
`that there are hash techniques that a POSITA could find and
`choose and use for this purpose. So he admitted that all of these
`components, all this technology was already out there. Nobody
`was improving a hash function. Nobody was improving OCR.
`Nobody was improving a barcode. All the technology was
`already out there and all they did was combine the generic
`
`
`
`
` 13
`
`
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
`PGR2020-00031
`Patent 10,445,966 B1
`
`computer components and generic computer functions and when
`you do that the case law is unambiguous. Those things are not
`eligible for patentable subject matter.
`
`When we look at the voter verified claim that was rejected
`and we're looking at slides 7 and 8, you see that the claim
`elements line up with all the claim elements of the claims here
`and that claim 1, I'm sorry claim 85 was found ineligible by the
`Federal Circuit. Now the pushback we heard from Patent Owner
`is that this was about human verification of the ballot, but what
`the Patent Owner admits at slide 9 where we have claim 1 of that
`same patent was also found invalid by the Federal Circuit and
`that included (c) and elements (c) and (d) in the second column
`there. The printer connected to said computer for printing the
`election ballot according to which the voter voted and then (d) a
`ballot scanning means for reading the votes on the printed ballot
`printed according to the election ballot which the voter voted
`that the votes shown on the printed ballot are compared by the
`computer program with the votes recorded in the computer for
`the voter. So here the voter votes in the computer, the computer
`prints out the ballot, the computer then scans the printout and
`compares that scan we submit by OCR, compares that scan of the
`ballot against its own computer record. None of that is done by
`humans. They simply -- so when you look at this, this claim
`element, but this was knocked out as well by the Federal Circuit
`under Section 101 and that's surely computer functionality to
`
`
`
`
` 14
`
`
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
`PGR2020-00031
`Patent 10,445,966 B1
`
`check a scan against an electronic record. There's no human
`cognition here at all in this claim element and this was found
`invalid under Section 101 as well.
`
`The other point I'd like to make is that in a letter to the
`Senate, the testimony from ES&S was that none of their current
`equipment had certain functionality. But that was just asking
`about their current equipment. That was a response to their
`current equipment. If you look at slide 22 at the time those
`statements were made, ES&S owned the Brockhouse patent,
`Exhibit 1004, and it says right in it that,
`"Tabulation device 14 may use optical character
`recognition (OCR) to decode the votes selections 114 on the
`printed activation cards 20."
`So the testimony to the Senate was none of our current
`equipment uses that OCR function but the intellectual property
`owned by ES&S at the time but not practiced yet certainly did
`disclose that function and so there's nothing more there than
`simply a statement before the Senate about how their equipment
`currently works and not a statement about what could or couldn't
`be done or what was possible or not possible. So unless you
`have more questions about Section 101 I will turn it over to Mr.
`Hartley for the prior art piece (phonetic).
`
`JUDGE REPKO: I do not. Do my colleagues have any
`questions?
`
`JUDGE WIEKER: No. Thank you.
`
`
`
`
` 15
`
`
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
`JUDGE BOUDREAU: Nothing further right now. Thank
`
`PGR2020-00031
`Patent 10,445,966 B1
`
`
`you.
`MR. EVANS: Thank you.
`
`MR. HARTLEY: Good afternoon. As Mr. Evans explained
`
`I'll be addressing grounds 2 to 7 of the petition. There's three
`patent overarching questions that arise in those grounds. The
`first is whether a person of ordinary skill in the art would have
`been motivated to use a dictionary containing potential voting
`choices when performing error correction on an OCR of a ballot
`which voting selections were printed or would the POSITA then
`just look to a generic English language dictionary that the
`POSITA knows does not have the voting selections, it doesn't
`have proper names, which makes common sense, and also the
`experts in this matter agreed that the person of ordinary skill in
`the art is going to use the dictionary that has the voting
`selections in it.
`
`The second overarching issue that we'll discuss today is
`whether Nadaf generates a cast vote record from the scanned
`OCR process in comparison of the OCR results to the barcode on
`the printed ballot in Nadaf. Again, the answer to that question is
`also yes. The same OCR data that is confirmed accurate by that
`barcode comparison has been counted after the audit is
`completed. That historic data is then confirmed accurate, it is a
`CVR it is a cast vote record. It is a record -- I'm sorry, I'm
`getting some echoing there. Somebody's not muted.
`
`
`
`
` 16
`
`
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
`PGR2020-00031
`Patent 10,445,966 B1
`
`JUDGE REPKO: Excuse me, can you please mute your
`
`phones. Thank you. You may proceed counsel.
`
`MR. HARTLEY: Okay. So returning back to issue No. 2
`and whether Nadaf presents the cast vote record. That data that's
`been OCR'd and is subsequently confirmed by the barcode
`comparison, the result there is used to count or record the vote.
`But that record is a record of the voter's voting selection. That
`is the definition of a cast vote record that's provided in the '966
`patent and that definition can be found at column 1, lines 23 to
`28.
`The third primary issue is whether a person of ordinary
`
`skill in the art would have been motivated to use the hash
`function with Nadaf to protect the integrity of the printed vote
`record. Again as Mr. Evans just explained, that's something
`that's been done in the voting industry for decades. So again the
`answer is yes, a person of ordinary skill and we're going to refer
`back repeatedly to the fact that this is a person of skill. They're
`not automatons. They're not forced to take disclosures from the
`prior art references and hold still and try to put them together
`like blocks of Legos. They're a person of skill in the art and in
`this instance the parties agree on what the definition of a person
`of ordinary skill in the art is.
`
`If we turn just briefly to slide 28. The parties here agree.
`A POSITA, a person of ordinary skill in the art, is going to have
`at least two years experience in voting machines, with electronic
`
`
`
`
` 17
`
`
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
`PGR2020-00031
`Patent 10,445,966 B1
`
`voting machines as well as all of the other components of the
`claims that we're going to be discussing today. Ballot marking
`devices, optical character recognition, barcode and hash
`functions, all of those features, all of those computer elements
`are something that the parties agree is within the skill of a
`person of ordinary skill in the art.
`
`So we first move to the first issue that we'll be addressing
`here and that's the use of an election dictionary and we can start
`there at slide 23. So here we have claims 1 and 15 that are
`subject to ground 2 and what we'll see here is the yellow claim
`elements that are common between them and there's really no
`dispute here that the yellow elements are disclosed in
`Brockhouse. That was the conclusion of the examiner during
`prosecution and the Applicant did not dispute that. Then we
`have the elements that are highlighted in red and those are the
`election dictionary elements and that's also important to note
`here, and this was pointed out at page 55 of the petition, the
`concept of an election dictionary is not new. It was disclosed in
`the Miyagawa reference that was cited during prosecution. The
`Applicant distinguished Miyagawa by arguing,
`"That Miyagawa makes no teaching of using the dictionary
`of potential vote choices as far as a process of confirming the
`accuracy of the OCR process itself."
`
`So (indiscernible) Miyagawa’s voter would write a voting
`selection and let's say write in their vote for Canada and they
`
`
`
`
` 18
`
`
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
`PGR2020-00031
`Patent 10,445,966 B1
`
`make a spelling error. So Miyagawa would have a list of correct
`voting selections and the system would compare them and oh,
`you misspelled George Washington. You substituted a B for an
`A or whatever they did. So again, the concept of an election
`dictionary was already known and as we're going to discuss as
`we going through here the fact that OCR errors occur was also
`known and the fact that dictionaries could be used to correct
`OCR errors was known. So again, nothing that's in these
`election dictionaries counts or elements of the claims was new.
`
`If we turn to slide 24. Here on the left we have the printed
`vote record that's disclosed in Brockhouse. At the top we have
`the barcodes. At the bottom in red we have the printed vote
`record. We have the voting selections that are printed on the
`printed vote record. The example highlighted in red, we have
`Best Vocal Artist. Those are for Frank Sinatra and Stevie Ray
`Vaughan and the voting selections are identified with reference
`No. 114. If we look at paragraph 56 of Brockhouse, that
`paragraph tells us that vote selections 114 are human readable
`text identifying each voting selection made by the voter, for
`example, General Motors and a brief description of the contest
`and/or candidate, e.g. Best Automobile Manufacturer. So
`Brockhouse teaches us they were scanning all the information
`that's on there, both the race for the contest and the vote
`selections that are made by the voter.
`
`On the right we have paragraph 134 from Brockhouse. It
`
`
`
`
` 19
`
`
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
`PGR2020-00031
`Patent 10,445,966 B1
`
`tells us that we're going to perform OCR and as a result of the
`OCR is what's going to be used in a vote tabulation, and that's
`the cast vote record. Again, there's no dispute here that those
`elements of the claims are met. There's also no dispute here that
`a person of ordinary skill in the art reading Brockhouse would
`know that the person of ordinary skill in the art is admittedly
`knowledgeable about OCR processes, that you're going to have
`errors in that OCR data set.
`
`If we turn then to slide 25, Dr. Lopresti, Patent Owner's
`expert confirms all of this. First piece of testimony highlighted
`on the left.
`
`"A. Yes, a POSITA would know that OCR could have
`errors."
`
`More specifically in the quote below. A person of ordinary
`skill in the art in July of 2018 would know using OCR in a ballot
`image they would expect to encounter errors and again in the
`middle quote, this is specifically addressing paragraph 134 of
`Brockhouse and the OCR process as disclosed in Brockhouse.
`They would realize -- a POSITA "would realize that there would
`likely be some errors." So again, nothing here is new. They
`noted there's going to be errors. We're talking about voting
`selections in a ballot or any ballot that's going to be used to
`officially tally the vote. A person of ordinary skill in the art
`knows you're going to get those vote selections right. You're
`going to correct errors if they exist, and lastly the quote here on
`
`
`
`
` 20
`
`
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
`PGR2020-00031
`Patent 10,445,966 B1
`
`slide 25 on the right. Dictionaries, Dr. Lopresti confirmed, were
`a known OCR correcting technique that predates the '966 patent.
`
`So then we look at slide 26. Here we have Heilper which is
`100 percent in line with what Dr. Lopresti has admitted a person
`of ordinary skill in the art would have known as of the filing
`date of the '966 patent. On the left in the Background of the
`Invention, Heilper tells us OCR steps are error prone. They're
`typically followed by an error correction step. For example, the
`computer may look up each OCR generated word in the
`dictionary and substitute the nearest match from the dictionary
`and here on the right Heilper tells us that these processes have
`been around since the 1960s.
`"Dictionary-based OCR error correction typically uses an
`approximate string-matching algorithm to find the nearest
`match."
`That's the exact same technology that's relied on in the '966
`patent. More specifically, we have the Levenshtein distance
`disclosed in 1966. That's the only specific algorithm that's
`disclosed in the '966 patent for performing any kind of error
`correction. Again, all of these elements long predate. Heilper
`was filed in December of 2003, so we're talking 15 years before
`the '966 patent was filed and just to highlight one more thing
`here, the quote on the left where we're substituting the nearest
`match from the dictionary. This is in the Background of the
`Invention of Heilper which is Exhibit 1005. You're substituting
`
`
`
`
` 21
`
`
`
`PGR2020-00031
`Patent 10,445,966 B1
`
`the nearest match from the dictionary which means there has to
`be a match in the dictionary. You're not going to match proper
`names to generic words in an English dictionary. A person of
`ordinary skill in the art would know that. The matches are going
`to be the potential voting selection. Again, this is all in the
`context of the OCR process in Heilper. It's not looking at
`Heilper in isolation without the context of Brockhouse.
`If we jump to slide 30 then. The ultimate question here on
`the validity of claims in ground 2 is would the person of ordinary
`skill in the art know to use a dictionary with the potential votin