`571-272-7822
`
`
`
`
`Paper # 46
`Entered: June 5, 2020
`
`
`UNITED STATES PATENT AND TRADEMARK OFFICE
`____________
`
`BEFORE THE PATENT TRIAL AND APPEAL BOARD
`____________
`
`CANON INC., CANON U.S.A., INC., and
`AXIS COMMUNICATIONS AB,
`Petitioner,
`
`v.
`
`AVIGILON FORTRESS CORPORATION
`Patent Owner.
`____________
`
`IPR2019-00311
`IPR2019-00314
`Patent 7,923,923 B2 & C1
`____________
`
`Record of Oral Hearing
`Held: April 8, 2020
`____________
`
`Before GEORGIANNA W. BRADEN, KIMBERLY McGRAW, and
`JESSICA C. KAISER, Administrative Patent Judges.
`
`
`
`
`
`
`IPR2019-00311
`IPR2019-00314
`Patent 7,923,923 B2 & C1
`
`
`APPEARANCES:
`
`ON BEHALF OF PETITIONER AXIS COMMUNICATIONS AB:
`
`
`C. GREGORY GRAMENOPOULOS, ESQUIRE
`Finnegan, Henderson, Farabow, Garrett & Dunner LLP
`901 New York Avenue, NW
`Washington, D.C. 20001-4413
`
`
`ON BEHALF OF PETITIONERS CANON INC. AND CANON U.S.A.
`INC.:
`
`
`RICHARD MARTINELLI, ESQUIRE
`Orrick Herrington & Sutcliffe
`1152 15th Street, NW
`Washington, D.C. 20005
`
`
`
`
`
`ON BEHALF OF THE PATENT OWNER:
`
`
`
`REZA DOKHANCHY, ESQUIRE
`Kirkland & Ellis
`1301 Pennsylvania Avenue, NW
`Washington, D.C. 20004
`
`
`The above-entitled matter came on for hearing on Wednesday,
`
`April 8, 2020, commencing at 1:10 p.m., via Video Teleconference.
`
`
`
`2
`
`
`
`IPR2019-00311
`IPR2019-00314
`Patent 7,923,923 B2 & C1
`
`
`P R O C E E D I N G S
`
`
`JUDGE MCGRAW: This is Judge McGraw and I think we are ready
`
`to begin. So, I would like to start by thanking everyone for your flexibility
`in conducting this remote hearing today. We realize this is a departure from
`our typical practice where we are able to have the parties and at least one
`judge together in the hearing room. Because this is a different type of
`procedure, I'd like to start by clarifying a few items.
`
`First, our primary concern is your right to be heard. If at any time
`during the hearing if you encounter technical or other difficulties that you
`feel fundamentally undermine your ability to adequately represent your
`client, please let us know immediately. For example, by contacting our team
`members who provided you with the connection information.
`
`Second, I'd like to remind everyone that there is no recording of this
`proceeding permitted. Therefore, no one is able to make an audio or a video
`recording. This is the same procedure that we would have in the actual court
`room hearing where nobody is allowed to bring in a recording device into
`the hearing.
`
`Third, when you're not speaking if you could please mute yourself that
`will help keep a clearer record. Fourth, to help the court reporter know who
`is speaking and to prepare an accurate transcript, we ask that everyone
`identify themselves before speaking and the judges, we will also attempt to
`do that each time that we ask a question.
`
`Also, we are very familiar with the record and we have access through
`our computer system to the entire record including the demonstratives.
`
`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
`
`
`
`3
`
`
`
`IPR2019-00311
`IPR2019-00314
`Patent 7,923,923 B2 & C1
`
`When you refer to the demonstratives, however, and you switch slides,
`please be sure to identify the new slide by page number so that we can
`follow along. Also, if you would like the panel to review a particular page
`or a document or exhibit in the record, after you identify that particular page,
`please pause for a few moments as it might take us a few seconds to navigate
`the record to get to that particular page in the record.
`
`Also counsel, because we are doing this remotely, you are not able to
`view the timer that we have in our hearing rooms. So, counsel are urged to
`use some sort of timer to keep track of their argument time. Later, I will get
`each party's rebuttal reserval time and we will make every effort to let you
`know when you are going into rebuttal and to give you a few minutes
`warning before all of your time is up. But please remember it's counsel's
`responsibility to keep track of time.
`
`Finally, please be aware that additional co-counsel, members of the
`public, including client representatives may be listening in to this oral
`hearing through an audio connection. So, please refrain from disclosing any
`confidential information. Does anyone have any questions about the issues
`that I just reviewed? Petitioner Mr. Martinelli?
`
`MR. MARTINELLI: No, Your Honor, that was very clear.
`
`JUDGE MCGRAW: Thank you and Mr., could you please pronounce
`your last name again?
`
`MR. DOKHANCHY: Sure, Your Honor, it's Dokhanchy.
`
`JUDGE MCGRAW: Thank you, Mr. Dokhanchy.
`
`MR. DOKHANCHY: Thank you, Your Honor and no questions from
`us.
`
`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
`
`
`
`4
`
`
`
`IPR2019-00311
`IPR2019-00314
`Patent 7,923,923 B2 & C1
`
`JUDGE MCGRAW: Thank you. As set forth in our oral hearing
`
`order and this is Judge McGraw speaking, each side has 75 minutes total to
`argue. Although the 311 and 314 [dropped audio 3:46] those cases will be
`argued together as they both involve the same patents and there is a
`substantial overlap of issues. A single transcript will be created an entered
`into each proceeding.
`Petitioner who has the burden of proof to show the unpatentability of
`the challenged claims by a preponderance of the evidence will begin by
`presenting its case. Patent Owner will then respond. Thereafter, Petitioner
`may use any time it has reserved for rebuttal to respond to Patent Owner's
`argument. Petitioner's rebuttal time may not be more than half of the
`allotted time. Thereafter, Patent Owner may present a brief surrebuttal.
`With that, I invite Petitioner to begin and how much time would you like to
`reserve for rebuttal?
`
`MR. MARTINELLI: Thank you, Your Honor. I'd like to reserve 35
`minutes.
`
`JUDGE MCGRAW: 35 minutes which means you have 40 minutes
`for your opening presentation.
`
`MR. MARTINELLI: All right, well I'll begin then.
`
`JUDGE MCGRAW: Thank you very much.
`
`MR. MARTINELLI: I thank all of you and all the staff for attending
`in these strange times that we're in and I appreciate everybody making
`themselves available. I'm Richard Martinelli. I will be representing Canon,
`Petitioner Canon in this hearing.
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`
`
`
`5
`
`
`
`IPR2019-00311
`IPR2019-00314
`Patent 7,923,923 B2 & C1
`
`I'd like to start with just a brief discussion of the 923 patent to orient
`
`ourselves before we jump into the issues. If you can turn to slide 4 of my
`demonstratives there's just some information giving a basic overview of the
`disclosure of the 923 patent. The 923 patent discloses -- lost the video there
`for a second, sorry everybody. The 923 patent discloses the concept of a
`system where video -- it's fundamentally for video retrieval. And the idea
`that you might want to be able to deal with video in a semantic sort of
`fashion.
`
`So, what the 923 talks about is having a process whereby video is
`analyzed to detect events, objects, attributes. Primarily the attributes of the
`objects are recorded and then later through a different process, a user could
`search the video attributes to try and find an interesting part of video. And
`the reason that the 923 states that this is important to do this this way is
`because if you wanted to look for a particular piece of video, you would
`have to actually maybe in prior art systems go through and scan the entire
`video.
`
`So, imagine if you're, you know, a security guard and you wanted to
`see if some event occurred, somebody went by some area of the system that
`you're monitoring or the site that you're monitoring. You would normally in
`an old school system have to go and just like manually review all of that
`video looking for whatever you're looking for. And the 923 says, wouldn’t
`it be better if we can define some attributes, dictate an event and then go and
`search just for that event.
`
`With that basic parameter out of the way, as was recognized in the
`institution decision, the prior art that we're dealing with here discloses
`
`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
`
`
`
`6
`
`
`
`IPR2019-00311
`IPR2019-00314
`Patent 7,923,923 B2 & C1
`
`exactly that concept. Both the Kellogg and Dimitrova references are
`directed to the idea that you want to look at the semantics of the video, you
`want to record semantic data about the video. You want to record attributes
`that give a user the freedom to then use that semantic data to go in and look
`for any particular piece of video that they might be interested in.
`
`And the institution decision found all of the limitations of the claim
`were reasonably shown at that stage and I think maybe we should spend
`most of our time here talking about what has the Patent Owner identified as
`its distinctions. And I think there is two big issues that I'd like to discuss.
`
`One is the fact that a number of these issues have already been
`decided in a prior IPR between the parties dealing with a related patent to
`the 923 patent. And as we'll discuss in more detail, the Board doesn't need
`to revisit and redecide issues that have already been before it and that have
`already been decided. The doctrine of collateral estoppel is available to
`meet the Board's burden of having to readdress issues that have already been
`addressed.
`
`Then even if we go to the substance of the particular issues and if we
`go to the substance of the issues that aren't subject to collateral estoppel, I
`think what we'll see is the distinctions the Patent Owner is making to try and
`overcome the prior art are all at a level of detail that's inconsistent with the
`specification of the 923 patent. I think what we have here is we have prior
`art references that are very detailed on how to actually implement the
`systems that they disclose. And a patent that has high level description and a
`Patent Owner that's criticizing the prior art (inaudible) details.
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`
`
`
`7
`
`
`
`IPR2019-00311
`IPR2019-00314
`Patent 7,923,923 B2 & C1
`
`I think there's a few places that we see this. One is with some of the
`
`constructions the Patent Owner is asking for that are sort of a distortion of
`the (inaudible) and when you look at many of them, they're not supported by
`the specification. Similarly, you know there's features of the prior art that
`are criticized by the Patent Owner but then the patent operates in exactly the
`same way. I think a good example there is trajectory attributes. The Patent
`Owner repeatedly criticizes the prior art for recording trajectory attributes.
`But the patent discloses exactly the same thing and tells you to record
`trajectory attributes.
`
`Then finally, there's features in the prior art that are described in some
`detail that frankly is lacking in the patent. The patent is higher level than the
`prior art is but the Patent Owner attempts to use that to criticize the prior art.
`And I think some examples there are the prior art's disclosure of indexing
`and the prior art's disclosure of the particular data structures it uses to store
`its data. So, you know, the Patent Owner tries to find distinction in these
`areas but there's just no support in the spec to distinguish the prior art at that
`granular level that the Patent Owner's attempting.
`
`JUDGE MCGRAW: And counselor are you still on slide 4 or have
`you moved on to slides 5, 6 and 7?
`
`MR. MARTINELLI: I have not moved on to any slides, I was just
`proceeding on.
`
`JUDGE MCGRAW: Very good.
`
`MR. MARTINELLI: And I'm guessing the way things are going, my
`things will not be super linear so I apologize for that. But actually, we can
`turn to slide 8 right now. And I'd like to briefly discuss some of the
`
`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
`
`
`
`8
`
`
`
`IPR2019-00311
`IPR2019-00314
`Patent 7,923,923 B2 & C1
`
`overarching basic concepts for collateral estoppel. Slide 8 shows the four
`main elements needed to show collateral estoppel. But I think the most
`important thing we need to sort of get a handle on is that collateral estoppel
`is about issue preclusion. So, it's tied to specific issues and it's a doctrine
`that says hey, if a prior tribunal addressed an issue, a future tribunal doesn't
`need to go back and revisit that same issue again.
`
`I think some of the criticisms that I've seen from Patent Owner relate
`to trying to turn it into a claim preclusion thing where they're trying to say
`we're attempting to completely tell the Board that they don't have to do
`anything here and they can just rely on the 661 decision. That's not true.
`Our collateral estoppel issues are tied to particular issues present in this case.
`
`So, if we go through the four elements of collateral estoppel, the first
`one is what I just said, right. Collateral estoppel is about issues and it's
`about showing that the issue in the current action is the same as the issue in
`the prior art.
`
`Number two, the issue was actually litigated in the first action. So,
`that was something where the issue was presented and the parties actually
`addressed the issue and I think that's true of all of the things we're asking for
`with respect to collateral estoppel.
`
`Three is similar but there's a nuanced difference here. It's that the
`final judgement in the first action was necessarily required for the
`determination of the identical issue. I don’t think that's a problem we have
`here but just to give you an example so you can get your head around what
`that is. In a typical patent litigation, you might have issues of infringement
`and the validity that are litigated by the parties. But at the end of the case,
`
`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
`
`
`
`9
`
`
`
`IPR2019-00311
`IPR2019-00314
`Patent 7,923,923 B2 & C1
`
`maybe the defendant shows that they're non-infringing. The parties might
`have hotly debated validity but the court never needs to reach that decision.
` So, in that case collateral estoppel wouldn’t apply. Here, the prior IPR
`was dealing with the exact same validity issues that we're dealing with so we
`don't have that problem. And then finally, the fourth element is that the
`parties or the party that the estoppel was going to apply against was fully
`represented. And again, I think Avigilon had adequate counsel and was
`fully represented in the 661 petition.
`
`If we can turn to slide 9. Slide 9 is a timeline that just shows the panel
`when collateral estoppel really attached here and when the different events
`occurred. So, the bottom half of this timeline is the timing associated with
`the 661 IPR. And you'll see that the final written petition for the 661 came
`in in March of 2019 that was after the petition was filed which was filed
`back in the petition here in the 923 patent the prior IPR was filed.
`
`Now, the actual final written decision typically wouldn’t be a -- or
`wouldn’t necessarily be a final determination. Because a Patent Owner
`could appeal and continue to keep those issues alive. And Avigilon did
`actually file a notice of appeal in July of 2019 but on October 2, 2019
`Avigilon withdrew the 661 IPR appeals. That was about a week before it
`provided its Patent Owner responses here in this proceeding. And at the
`point that Avigilon withdrew that appeal, the final determination of the 661
`panel became completely final.
`
`Another interesting while we're just talking briefly about the 661
`patent. I think it's important to know how the 661 patent is related to the
`923 patent. Because sometimes issues between related patents could get
`
`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
`
`
`
`10
`
`
`
`IPR2019-00311
`IPR2019-00314
`Patent 7,923,923 B2 & C1
`
`changed based on the disclosures. So, the 661 patent is, I don’t know how
`many greats it is but it's a great, great, great grandniece or nephew to the 923
`patent with some intervening (inaudible). I think the critical thing to note is
`that the disclosure provided in the 661 patent was more extensive.
`
`So, hypothetically the 661 patent might have more detail that would
`allow somebody to raise a distinction and to distinguish over the prior art.
`But what we have here is the 923 patent has less detail. So, everything that's
`relevant in the 661 patent to deciding that determination that the 661 patent
`was validated, that narrow patent, it's even broader for the 923.
`
`So, if we proceed to slide 10 and we'll talk about some of these in
`more detail as we get to the different issues. This is just a short sort of
`outline of some of the things that were decided in the 661, heavily litigated
`subject to a fully reasoned opinion. That we think Avigilon is now
`continuing to pursue and that the Board really shouldn't be burdened to
`revisit.
`
`The first issue is the construction of user rule and whether a user rule
`requires a response or not. That exact issue was addressed in full in the 661
`IPR and admittedly here they call it a new user rule. But there's no argument
`that the addition of the word new fundamentally changes what the user rule
`is.
`The second issue which is a big one that we'll talk about is this
`
`independence concept. And again, that goes back to the overview we had of
`what the concept is here. The concept for the 923 is detecting attributes and
`then later defining an event in the context of those attributes. Now that was
`something that again was front and center in the 661, that exact same issue
`
`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
`
`
`
`11
`
`
`
`IPR2019-00311
`IPR2019-00314
`Patent 7,923,923 B2 & C1
`
`and the prior panel found that the prior art teaches that limitation and
`construed the claim as necessary to find that showing.
`
`Next getting into the actual prior art itself there was the ruling that
`both that Kellogg teaches the independence elements. There was a ruling
`that the prior art references that we're combining here, Kellogg and Brill
`would be combined by a person of ordinary skill in the art. And then the
`same for Dimitrova with its independence teaching and Dimitrova with its
`combination.
`
`So, I think that's it for the overview of the collateral estoppel. I was
`going to get into the issue of attributes and events next. Of course, if there's
`anything that the panel wants to hear specifically just let me know.
`
`JUDGE MCGRAW: We are familiar with the record so feel free to
`move along. Thank you.
`
`MR. MARTINELLI: Absolutely. So, with respect to the distinction
`between attributes and events, if you look at slide 12, Patent Owner, and I'm
`doing this first in the context of Kellogg. A lot of the issues are the same
`and we can talk about them in the context of Dimitrova but we'll start with
`Kellogg. When you go through the Patent Owner's responses, I believe
`there's a number of different places where they fundamentally criticize the
`prior art because they say it's detecting events and not attributes. And they
`make that argument for a number of different limitations and so slide 12 just
`sort of collects the different limitations where we saw the Patent Owner
`making the same fundamental argument.
`
`When we get to -- if we can move onto slide 13. Slide 13 lays out just
`the basic intro to Kellogg and what Kellogg is doing. So, Kellogg was the
`
`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
`
`
`
`12
`
`
`
`IPR2019-00311
`IPR2019-00314
`Patent 7,923,923 B2 & C1
`
`thesis of an MIT student and what Kellogg was trying to accomplish was to
`provide a really robust way to store data about video and then to make that
`data about the video retrievable later. And a quote here that I'd like to point
`out, visual memory could act as an interface between inputs and
`applications. So, inputs are the video that you want to analyze and
`applications in a computer system, those are the, you know, actual semantic
`part that wants to look at the video and do some work with it.
`
`And they give an example, for example, computer vision algorithms
`for a security system could analyze data provided by various cameras and
`store information in the visual memory. Applications could then retrieve the
`data to track objects, watch for suspicious events and respond to user
`queries. The visual memory would coordinate this information and
`eliminate the need for full connectivity between the inputs and the outputs.
`
`Again, that's a hallmark of independence here, right? Like they're
`eliminating and they're putting an abstraction layer between the part of the
`system that does the video and the part of the system that actually tries to,
`you know, get the results it wants out of the video.
`
`If we proceed to slide 14, I have here an example of one of the events
`that Kellogg discloses could be searched for. And so, this is part of the
`queryng and the syntax that Kellogg provides. And you'll see this is an ad
`hoc thing. Cutting through some of this code, essentially what it says is you
`have a variable person p, you have a timeline range and then you have a
`second variable person q. And then you do the heart of the work which is
`look for situations where person q comes within three units of person p.
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`
`
`
`13
`
`
`
`IPR2019-00311
`IPR2019-00314
`Patent 7,923,923 B2 & C1
`
`So, that is something that a user can do in Kellogg. It's completely ad
`
`hoc and independent from the data that was collected. For example, if I
`came along and wanted to, you know, look for a closer connection between
`the centroid I could write this query and put in a smaller number. If I later
`wanted to search for an intersection of three people, I can frame a query to
`add a third person and put in a different dimension. This is something that's
`completely unknown at the point of the video detection part of Kellogg.
`
`So, if we proceed on to slide 15, slide 15 through 17, I don't know that
`I need to belabor each of them. But they basically collect a variety of
`statements from the 661 petition showing that this issue was directly
`addressed. And the prior panel found that Kellogg practices the independent
`that was described in the patent.
`
`If we go to slide 18, we have the claim construction issue around
`independence. And this is (dropped audio).
`
`JUDGE MCGRAW: I believe that Patent Owner, this is Judge
`McGraw, I believe that Patent Owner has an argument about an issue that
`was not present in the 661 IPR proceedings. And that is Kellogg does not
`teach the “only” limitation as the term “only” was not present in the 661
`claims. And Patent Owner when it's their turn to argue can correct me if I'm
`wrong.
` But I understand their argument to be that when Kellogg performs the
`searches that you're identifying here, such as the approach query, that it is
`not performed on the bucket list index. And because of that the Kellogg
`visual system does not satisfy the “only” limitation that's present in the
`claims at issue in this proceeding.
`
`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
`
`
`
`14
`
`
`
`IPR2019-00311
`IPR2019-00314
`Patent 7,923,923 B2 & C1
`
`MR. MARTINELLI: Okay yes, Judge McGraw, I think that is an
`
`issue and there's a couple points there. One I think is a claim construction
`issue that we'll have to address and then another is what are the fundamental
`teachings of Kellogg. And I'll do the Kellogg teachings first if we could go
`to slide 34 in my presentation.
`
`So, what we rely on to show that Kellogg is searching only the
`attributes is this bucket index that's disclosed in Kellogg. The bucket index
`is basically a quick and dirty plain way of storing the attributes. And the
`way Kellogg describes it is that you'll collect the attributes and you'll just --
`it's like you're throwing them a bucket. You'll just put them in a list that you
`can go back and search later.
`
`Now that raw set of attributes can be searched. If I understand Patent
`Owner's argument with respect to this, they're somehow saying it's in the
`wrong section of Kellogg or it wasn't, you know, the system of Kellogg.
`Well, Kellogg has a bunch of embodiments and the bucket index is one of
`the embodiments. And, in fact, Kellogg actually built the system using the
`bucket index and used it to compare to other ways of organizing the data that
`it taught.
`
`So, if you look on slide 34, there's a nice signal that they have where
`they compare the performance of the bucket index to to the quadtree and
`grid indexes. Now, you know, Kellogg was trying to show some fancy
`technology with the quadtree and grid indexes and they did have advantages
`but that doesn't mean that the bucket index wasn't a potential embodiment.
`And, in fact, Kellogg actually teaches some advantages of the bucket index.
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`
`
`
`15
`
`
`
`IPR2019-00311
`IPR2019-00314
`Patent 7,923,923 B2 & C1
`
`So, if we look at page 83 of Kellogg and I'll just read it, I don’t know
`
`if you have to actually pull it up. The bucket index simply maintains a list of
`all the objects stored in the visual memory. Since there are no overheads for
`storage of complex index structures, a bucket index can achieve the highest
`update performance.
`
`So, if you had an embodiment and you were a person of skill in the art
`who wanted to build something where update performance was important,
`Kellogg would tell you, use a bucket index. So, I don't see anything here
`that suggest that the bucket index is not a valid teaching and not a valid
`embodiment. And again, tying it to the approach query, this would apply to
`all the queries. This is a situation where the teachings of Kellogg are
`optional elements that would make sense to be added into any of the
`different embodiments. So, I don't think there's an issue --
`
`JUDGE KAISER: So, Mr. Martinelli, this is Judge Kaiser. I just
`want to pause for a quick second and I apologize for interrupting you. I
`don't see anybody's video now so I just want to make sure that you all can
`hear me and that we still have everybody.
`
`MR. MARTINELLI: So, I can hear you, Judge Kaiser.
`
`MR. DOKHANCHY: This is Reza Dokhanchy. I can hear you as
`well, Judge Kaiser.
`
`
`
`JUDGE KAISER: Okay, all right. Well we'll just sort of proceed on
`the basis that at least we still have the audio. But I encourage anyone to
`raise their hand or to call the alternate number if we have this degrade
`further. And maybe I'll just ask since we are experiencing some technical
`
`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
`
`
`
`16
`
`
`
`IPR2019-00311
`IPR2019-00314
`Patent 7,923,923 B2 & C1
`
`difficulty at least on the video. Counsel, you both received an alternative
`call in number if we have technical difficulties, is that right?
`
`MR. MARTINELLI: Yes, Your Honor.
`
`JUDGE KAISER: Okay. So, I do have a question and sorry for the
`side track into the technical issues.
`
`MR. MARTINELLI: No worries.
`
`JUDGE KAISER: I think we're all doing our best. So, I want to
`make sure I understand your argument, Mr. Martinelli, around what you
`referred to as the various or the different embodiments in Kellogg. Because
`you do have an anticipation ground based on Kellogg. And so, I'm trying to
`understand if your position is that it's okay for you to rely on the different
`embodiments in Kellogg as part of that or if what you're saying is essentially
`there are a bunch of different options in Kellogg that apply sort of regardless
`of the embodiment so sort of a mix and match is okay.
`
`MR. MARTINELLI: Right. So, I think the position is that Kellogg
`teaches a person of skill in the art in these embodiments to use each of these
`things and when to use them. So, if you look at some of the cases where
`this, you know, issue of anticipation comes in, I think it's a very different
`situation than what we have in Kellogg. I think what you have in those
`cases is situations where there's a holistic part. You know, I think one of the
`cases dealt with protocols. And you've got two independent alternatives, say
`two protocols and the district court in that case erred because it took
`subsections of one protocol and subsections of another protocol and mixed
`and matched those together.
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`
`
`
`17
`
`
`
`IPR2019-00311
`IPR2019-00314
`Patent 7,923,923 B2 & C1
`
`So, you know, if you were imagining say somebody had a patent
`
`where they disclosed you could use WiFi or Bluetooth and then gave some
`detail about elements of WiFi and elements of Bluetooth. I think it would be
`improper to say I'm going to make an anticipation where I create this new
`hybrid protocol that's neither WiFi nor Bluetooth but some amalgam of the
`two. That's not what we have here.
`
`What we have here is a situation where Kellogg says you can use a
`bucket index, you can use a quadtree, you can use any other index you want
`and all of them are good and useable for their case. So, I'm not taking, you
`know, things that are unique to the bucket index and applying them across to
`something that Kellogg distinguishes. So, the searching that you do on the
`index or on the set of attributes is unrelated to the way you organize those
`attributes.
`
`And again, I think this is another situation where the Patent Owner
`has sort of gotten away from what the 923 patent actually teaches. So,
`there's no basis in the 923 patent to support defined distinctions over how
`you store data. There's actually precious little to support Patent Owner's
`only construction that they're advocating here. And I'm sure the Patent
`Owner will spend some time on this so I don’t want to make their arguments
`for them.
`
`But the only limitation I think they're arguing should be given to plain
`and ordinary meaning and I think the prior art shows “only” under the plain
`and ordinary meaning. The Patent Owner then, despite saying it's the plain
`and ordinary meaning, adds in a bunch of concepts about abstractions that I
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`
`
`
`18
`
`
`
`IPR2019-00311
`IPR2019-00314
`Patent 7,923,923 B2 & C1
`
`find no support for in the specification at all. And I've got a couple slides on
`this.
`So, slide 30 and 31 in my presentation has every reference to the term
`
`abstraction in the 923 patent with the exception of the word abstract on the
`front page of the patent. There's nothing in here that distinguishes
`abstractions and attributes in the way that Patent Owner is attempting to use
`them to try and distinguish the prior art. There's nothing in here about the
`type of indexing you need to use, you know, distinguishing one type of
`indexing from another. There's nothing in here about how you store the
`data, it's just simply not in the specification and not supportable.
`
`I could read a couple of these. So, on slide 30, we've got the
`statement, in block 31, one or more object types of interest are identified in
`terms of video primitives or abstractions thereof. Okay, fair enough.
`Examples of one or more objects include an object, a person, a red object,
`two objects, two persons, a vehicle. So, nothing in there even tells you how
`you distinguish between the video primitives and the abstraction must less,
`you know, the extra gloss that Patent Owner is trying to put on abstractions
`in its only construction.
`
`Then in the next paragraph in block 35, one or more discriminators
`are identified by describing interactions between video primitives or their
`abstractions, spatial areas of interest and temporal attributes of interest.
`Again, nothing in there to say indexes are bad or, you know, what an
`abstraction is or how you distinguish between abstractions or not
`abstractions.
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`
`
`
`19
`
`
`
`IPR2019-00311
`IPR20