Paper No. 28
`
`
`
`
`
`
`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.
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`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
Accessing this document will incur an additional charge of $.
After purchase, you can access this document again without charge.
Accept $ Charge
Still Working On It
This document is taking longer than usual to download. This can happen if we need to contact the court directly to obtain the document and their servers are running slowly.
Give it another minute or two to complete, and then try the refresh button.
A few More Minutes ... Still Working
It can take up to 5 minutes for us to download a document if the court servers are running slowly.
Thank you for your continued patience.
This document could not be displayed.
We could not find this document within its docket. Please go back to the docket page and check the link. If that does not work, go back to the docket and refresh it to pull the newest information.
Your account does not support viewing this document.
You need a Paid Account to view this document. Click here to change your account type.
Your account does not support viewing this document.
Set your membership
status to view this document.
With a Docket Alarm membership, you'll
get a whole lot more, including:
- Up-to-date information for this case.
- Email alerts whenever there is an update.
- Full text search for other cases.
- Get email alerts whenever a new case matches your search.
One Moment Please
The filing “” is large (MB) and is being downloaded.
Please refresh this page in a few minutes to see if the filing has been downloaded. The filing will also be emailed to you when the download completes.
Your document is on its way!
If you do not receive the document in five minutes, contact support at support@docketalarm.com.
Sealed Document
We are unable to display this document, it may be under a court ordered seal.
If you have proper credentials to access the file, you may proceed directly to the court's system using your government issued username and password.
Access Government Site
