`
`2
`
`3
`
`4
`
`5
`
`6
`
`7
`
`8
`
`9
`
`10
`
`11
`
`12
`
`13
`
`14
`
`15
`
`16
`
`17
`
`18
`
`19
`
`20
`
`21
`
`22
`
`23
`
`24
`
`25
`
` UNITED STATES PATENT AND TRADEMARK OFFICE
`
` BEFORE THE PATENT TRIAL AND APPEAL BOARD
`
`- - - - - - - - - - - - - - - - -x
`
`CANON INC., CANON U.S.A., INC., :
`
`and AXIS COMMUNICATIONS AB,
`
` :
`
` Petitioner,
`
`: CASE NO'S.
`
`v.
`
`: IPR2019-00311
`
`AVIGILON FORTRESS CORPORATION,
`
`: IPR2019-00314
`
` Patent owner.
`
`: PATENT NO.
`7,932,923
`
`- - - - - - - - - - - - - - - - -x
`
` Videotaped Deposition of JOHN GRINDON, D.Sc
`
` Washington, D.C.
`
`Wednesday, October 2, 2019
`
` 9:36 a.m.
`
`Reported by:
`
`Cassandra E. Ellis, RPR
`
`Job No.: 10060621
`
`AVIGILON EX. 2018
`IPR2019-00314
`Page 1 of 187
`
`
`
`1 Videotaped Deposition of JOHN GRINDON, D.Sc, held at the
`
`2 offices of Kirkland & Ellis, LLP, 1301 Pennsylvania
`
`3 Avenue, Northwest, Washington, D.C. 20004, pursuant
`
`4 to agreement, before Cassandra E. Ellis, Certified
`
`5 Court Reporter - Virginia, Certified Court Reporter -
`
`6 Washington, Certified Shorthand Reporter - Hawaii,
`
`7 Registered Professional Reporter, Certified Livenote
`
`8 Reporter, Realtime Systems Administrator, Registered
`
`9 Professional Reporter, and Notary Public of The
`
`10 District of Columbia.
`
`11
`
`12
`
`13
`
`14
`
`15
`
`16
`
`17
`
`18
`
`19
`
`20
`
`21
`
`22
`
`23
`
`24
`
`25
`
`AVIGILON EX. 2018
`IPR2019-00314
`Page 2 of 187
`
`
`
`1 A P P E A R A N C E S
`
`2 ON BEHALF OF PETITIONER CANON:
`
`3 RICHARD MARTINELLI, ESQUIRE
`
`4 ORRICK, HERRINGTON & SUTCLIFFE, LLP
`
`5 51 West 52nd Street
`
`6 New York, New York 10019
`
`7 (212) 506-3702
`
`8 Rmartinelli@orrick.com
`
`9 ON BEHALF OF PETITIONER AXIS:
`
`10 GUANG-YU ZHU, ESQUIRE
`
`11 FINNEGAN, HENDERSON, FARABOW,
`
`12 GARRETT & DUNNER, LLP
`
`13 901 New York Avenue, Northwest
`
`14 Washington, D.C. 20001
`
`15 (202) 408-4000
`
`16 Guang-yu.zhu@finnegan.com
`
`17 ON BEHALF OF PATENT OWNER:
`
`18 REZA DOKHANCHY, ESQUIRE
`
`19 KIRKLAND & ELLIS LLP
`
`20 555 California Street
`
`21 San Francisco, California 94104
`
`22 (415) 439-1469
`
`23 Reza.dokhanchy@kirkland.com
`
`24
`
`25 ALSO PRESENT: Robyn A. Ellis, Legal Videographer
`
`AVIGILON EX. 2018
`IPR2019-00314
`Page 3 of 187
`
`
`
`1 C O N T E N T S
`
`2 EXAMINATION OF JOHN GRINDON, D.Sc PAGE
`
`3 By Mr. Dokhanchy 6
`
` E X H I B I T S
`
`4 5 6 7
`
`8 (Attached to the Transcript)
`
`9 JOHN GRINDON, D.Sc Deposition Exhibit PAGE
`
`10 Exhibit 1001 US Patent 7,932,923 B2 46
`
`11 Exhibit 1003 Visual Memory by Christopher 13
`
`12 James Kellogg
`
`13 Exhibit 1006 Transactions Information Systems 129
`
`14 Special Issue on Video Information Retrieval
`
`15 Exhibit 1016 Video Surveillance System 102
`
`16 Employing Video Primitives Amendment and
`
`17 Reply
`
`18 Exhibit 1022 Object-Oriented Conceptual 122
`
`19 Modeling of Video Data
`
`20
`
`21
`
`22
`
`23
`
`24
`
`25
`
`AVIGILON EX. 2018
`IPR2019-00314
`Page 4 of 187
`
`
`
`1 P R O C E E D I N G S
`
`2 THE VIDEOGRAPHER: Good
`
`3 morning. This is the beginning of
`
`4 disk number one in the deposition
`
`5 of Dr. John Grindon, taken in the
`
`6 matter of Canon, Inc., Canon USA,
`
`7 and Axis Communications AB,
`
`8 petitioner, versus Avigilon
`
`9 Fortress Corporation, patent
`
`10 owner, Case Number IPR2019-00311
`
`11 and IPR2019-00314, held in the
`
`12 United States Patent and Trademark
`
`13 Office before the Patent Trial and
`
`14 Appeal Board.
`
`15 Today's date is October 2nd,
`
`16 2019, and the time on the monitor
`
`17 is 9:36 a.m. My name is Robyn
`
`18 Ellis, and I am the legal
`
`19 videographer, the court reporter
`
`20 is Cassandra Ellis. We are
`
`21 representing Aptus Court
`
`22 Reporting.
`
`23 Counsel appearances will be
`
`24 noted on the stenographic record.
`
`25 Will the court reporter
`
`AVIGILON EX. 2018
`IPR2019-00314
`Page 5 of 187
`
`
`
`1 please swear in the witness then
`
`2 you may proceed.
`
`3 JOHN GRINDON, D.Sc
`
`4 having been sworn, testified as follows:
`
`5 EXAMINATION
`
`6 BY MR. DOKHANCHY:
`
`7 Q Good morning, Dr. Grindon.
`
`8 A Good morning.
`
`9 Q You've been deposed several
`
`10 times before; is that right?
`
`11 A Yes.
`
`12 Q Can you approximate how many
`
`13 times?
`
`14 A I don't have an exact count. I
`
`15 would say 15 to 20, something like that.
`
`16 Q Okay. So I'm sure you're
`
`17 familiar with the basic mechanics of a
`
`18 deposition, but if anything is unclear
`
`19 about the procedure or about any of my
`
`20 questions, just feel free to let me know,
`
`21 okay?
`
`22 A Will do.
`
`23 Q All right. I'll just ask that
`
`24 we not talk over each other, so that the
`
`25 court reporter can get everything down,
`
`AVIGILON EX. 2018
`IPR2019-00314
`Page 6 of 187
`
`
`
`1 is that all right?
`
`2 A Sure.
`
`3 Q Is there any reason you cannot
`
`4 give truthful and accurate testimony
`
`5 today?
`
`6 A No.
`
`7 Q Could you please state and
`
`8 spell your full name, for the record?
`
`9 A Yes, John Grindon, J-o-h-n,
`
`10 G-r-i-n-d-o-n.
`
`11 Q Okay. Dr. Grindon, you
`
`12 understand that you're here to testify in
`
`13 connection with US patent number 9 --
`
`14 excuse me -- 7,932,923; is that right?
`
`15 A Yes.
`
`16 Q And if I call that the `923
`
`17 patent, you'll know what I'm talking
`
`18 about?
`
`19 A Yes.
`
`20 Q And you understand that the two
`
`21 IPR proceedings you're here to testify
`
`22 about are IPR2019-311 and -314; correct?
`
`23 A Yes.
`
`24 Q And if I call those the 311 and
`
`25 314 proceedings you'll know what I'm
`
`AVIGILON EX. 2018
`IPR2019-00314
`Page 7 of 187
`
`
`
`1 talking about; right?
`
`2 A Yes.
`
`3 Q Okay. When were you retained
`
`4 to work on the 311 and 314 IPRs?
`
`5 A I don't have an exact date.
`
`6 It's been something over a year, I
`
`7 believe.
`
`8 Q Okay. When did you first
`
`9 review the `923 patent?
`
`10 A Again, I don't have an exact
`
`11 date, but it's been something over a
`
`12 year.
`
`13 Q Okay. Had you reviewed the
`
`14 `923 patent before you were retained to
`
`15 work on these two IPR matters?
`
`16 A I -- I may have, I'd have to go
`
`17 back and refresh my memory on that.
`
`18 Q Okay.
`
`19 A I don't recall right now.
`
`20 Q Okay. And I see from your CV
`
`21 you have a substantial engineering
`
`22 background; were any of your jobs
`
`23 specific to surveillance systems?
`
`24 A Specific to surveillance?
`
`25 Q Correct.
`
`AVIGILON EX. 2018
`IPR2019-00314
`Page 8 of 187
`
`
`
`1 A I would say yes. Now, when I
`
`2 worked with McDonnell Douglas I was
`
`3 working on the cruise missile program,
`
`4 and part of the work there was to develop
`
`5 a video system to scan the terrain and
`
`6 determine the position of the missile
`
`7 relative to objects that it found on the
`
`8 terrain and so forth.
`
`9 Q Okay. Did that involve
`
`10 recognizing human activity in any way?
`
`11 A That particular one -- could
`
`12 you tell me what you mean by human
`
`13 activity?
`
`14 Q I don't mean anything other
`
`15 than the ordinary usage of the words,
`
`16 "human activity."
`
`17 A Okay. I don't see human
`
`18 activity in the claims, that's why I'm
`
`19 asking what your meaning is of that term.
`
`20 Also, as part of that work, I developed
`
`21 algorithms for determining and
`
`22 recognizing targets, such as ships at
`
`23 sea. So this would involve human
`
`24 activity, the navigation of the ships.
`
`25 I also worked on a missile
`
`AVIGILON EX. 2018
`IPR2019-00314
`Page 9 of 187
`
`
`
`1 system, using laser radar, known as
`
`2 LIDAR, L-I-D-A-R, and again, developed
`
`3 algorithms to detect human-constructed
`
`4 objects in the scene.
`
`5 Q And so you agree that the `923
`
`6 patent concerns recognizing objects and
`
`7 then later analyzing those objects; do
`
`8 you agree with that general premise?
`
`9 MR. MARTINELLI: Objection
`
`10 to the form.
`
`11 A The -- the general premise, did
`
`12 you -- can you state that again, please?
`
`13 Q Sure. The `923 patent concerns
`
`14 detecting objects, including attributes
`
`15 of the objects, and then later analyzing
`
`16 those attributes to determine events; do
`
`17 you agree with that?
`
`18 A When you say: "Later
`
`19 analyzing," what do you -- what do you
`
`20 mean by analyzing? What do you include
`
`21 in that?
`
`22 Q I just mean however you
`
`23 understand the term analyzing.
`
`24 A Well, what I would prefer to do
`
`25 is maybe to look at the patents and --
`
`AVIGILON EX. 2018
`IPR2019-00314
`Page 10 of 187
`
`
`
`1 and see what the patent says.
`
`2 Q Okay. So without looking at
`
`3 the patent you can't agree or disagree
`
`4 with my statement; is that right?
`
`5 A The -- the patents analyzes
`
`6 video to derive attributes and then,
`
`7 later, there are rules applied to the
`
`8 attributes to detect events. If
`
`9 that's -- if that's what you were saying,
`
`10 then I would agree.
`
`11 Q Okay. Does any of your
`
`12 experience in the engineering field
`
`13 relate to detecting attributes from a
`
`14 video and then applying rules to those
`
`15 attributes to detect events?
`
`16 A Yes.
`
`17 Q Okay. Can you please explain?
`
`18 A This would be some of the
`
`19 research, again, at McDonnell Douglas,
`
`20 among other things. But in developing
`
`21 algorithms and techniques for recognizing
`
`22 objects from video imagery, for the
`
`23 purposes of cruise missile guidance, and
`
`24 then determining, for example, after
`
`25 detecting an object, as I mentioned a
`
`AVIGILON EX. 2018
`IPR2019-00314
`Page 11 of 187
`
`
`
`1 moment ago, a shipwreck ignition,
`
`2 automatic target recognition, extracting
`
`3 measurements from the object and going
`
`4 through a classification process to
`
`5 determine whether this was a, for
`
`6 example, a friend or foe ship. So that's
`
`7 one example.
`
`8 Q Okay. In that context, what
`
`9 were the events that would have been
`
`10 determined based on applying rules?
`
`11 MR. MARTINELLI: Objection,
`
`12 form.
`
`13 A The -- the video would be
`
`14 analyzed to determine the attributes,
`
`15 define objects and determine the
`
`16 attributes of the object, and then to --
`
`17 in the -- in the case of the ship
`
`18 example, the appearance of a -- of a
`
`19 particular type of ship in the video
`
`20 would be one event that was determined,
`
`21 the time that it appeared, so the -- the
`
`22 appearance of a particular type of object
`
`23 at a particular time would be one
`
`24 example.
`
`25 Q Okay. One of the references
`
`AVIGILON EX. 2018
`IPR2019-00314
`Page 12 of 187
`
`
`
`1 that you analyzed in your declaration
`
`2 that you submitted in connection with the
`
`3 311 and 314 IPRs is called the Kellogg
`
`4 reference; is that right?
`
`5 A Yes.
`
`6 MR. DOKHANCHY: All right.
`
`7 Let me hand you a copy of that.
`
`8 And this is Exhibit 1003, the
`
`9 Kellogg reference.
`
`10 (Previously marked Exhibit
`
`11 No. 1003 was identified for the
`
`12 record.)
`
`13 BY MR. DOKHANCHY:
`
`14 Q And this is the reference that
`
`15 you opined on in your declaration; is
`
`16 that right?
`
`17 A This appears to be the one,
`
`18 yes.
`
`19 Q Okay. And Kellogg uses an
`
`20 object-oriented database; correct?
`
`21 A That's a fair characterization.
`
`22 Q What is your understanding of
`
`23 what an object-oriented database is?
`
`24 A I think what we should do is
`
`25 take a look at, specifically, what
`
`AVIGILON EX. 2018
`IPR2019-00314
`Page 13 of 187
`
`
`
`1 Kellogg is describing as his particular
`
`2 database.
`
`3 Q Okay. And we can do that, but
`
`4 first I want to get your understanding.
`
`5 Do you have an understanding of what an
`
`6 object-oriented database is?
`
`7 A Again, I would prefer to look
`
`8 at Kellogg's understanding and
`
`9 specification of what he puts in the
`
`10 database.
`
`11 Q Okay. Can you answer my
`
`12 question: Do you have an understanding
`
`13 of what an object-oriented database is?
`
`14 A An object-oriented database
`
`15 would have to do with the object or the
`
`16 particular -- the particular object, as
`
`17 opposed to the identifier of -- let's
`
`18 see -- I'm having trouble coming up with
`
`19 a concise definition. I don't know if I
`
`20 defined that in my declaration or not.
`
`21 We could take a look. Why don't we do
`
`22 that?
`
`23 Q Okay. So you want to look at
`
`24 your declaration for a definition of
`
`25 object-oriented database; is that what
`
`AVIGILON EX. 2018
`IPR2019-00314
`Page 14 of 187
`
`
`
`1 you're doing?
`
`2 A I don't know that I defined it
`
`3 in there, but I'm thinking of how Kellogg
`
`4 is using the -- the visual database.
`
`5 Q Okay. I'm not sure that you
`
`6 provided a definition, but I guess
`
`7 without flipping through your
`
`8 declaration, which is several hundred
`
`9 pages, is there anything you can tell me
`
`10 about how an object-oriented database
`
`11 works?
`
`12 A In a database that is organized
`
`13 by the content or subject of a -- for
`
`14 example, in this case, a video, which
`
`15 then would give you the attributes and
`
`16 other characteristics of the object,
`
`17 would be a characterization.
`
`18 Q Okay. And with respect to
`
`19 those items, the attributes and
`
`20 characteristics of the object are
`
`21 structured around the object itself;
`
`22 right?
`
`23 MR. MARTINELLI: Objection,
`
`24 form.
`
`25 A That sounds like a fair
`
`AVIGILON EX. 2018
`IPR2019-00314
`Page 15 of 187
`
`
`
`1 statement.
`
`2 Q And the way that you would
`
`3 access those attributes are to access
`
`4 them through the object data structure;
`
`5 right?
`
`6 A That's essentially right.
`
`7 Q Okay. So in Kellogg there are
`
`8 a number of objects described, and we can
`
`9 feel free to look at them, starting on
`
`10 page 22. There are three main types of
`
`11 objects in Kellogg, that is a spatial
`
`12 object, a temporal object, and a
`
`13 spatiotemporal object; correct?
`
`14 A Yes.
`
`15 Q And, for example, a
`
`16 spatiotemporal object is what you would
`
`17 use to represent a moving object or a
`
`18 moving person; right?
`
`19 A You could use that to describe
`
`20 that, so that would be the joint spatial
`
`21 temporal aspects of an object as opposed
`
`22 to each one independently.
`
`23 Q And so, for example, the
`
`24 spatiotemporal object is what you would
`
`25 use to represent a persing -- person
`
`AVIGILON EX. 2018
`IPR2019-00314
`Page 16 of 187
`
`
`
`1 walking down a hall; right?
`
`2 A That -- in general, yes.
`
`3 Q And turning back to the spatial
`
`4 object, the spatial object in Kellogg is
`
`5 defined by a set of points; correct?
`
`6 A Can you -- I don't know that I
`
`7 would just limit it to a set of points.
`
`8 Can you show me where you're looking?
`
`9 Q Sure. If you turn to page 24,
`
`10 if you go to the second to last
`
`11 paragraph, it says: A spatial object is
`
`12 defined by a set of points and a local
`
`13 coordinate system. This information is
`
`14 sufficient to fully represent a spatial
`
`15 object; do you see that?
`
`16 A Yes.
`
`17 Q Okay. And so you agree that in
`
`18 Kellogg a spatial object is defined by a
`
`19 set of points and a local coordinate
`
`20 system; right?
`
`21 A And a local coordinate system.
`
`22 Q And Kellogg says that
`
`23 additional information, such as size,
`
`24 measurements, and other factors, besides
`
`25 the set of points and local coordinate
`
`AVIGILON EX. 2018
`IPR2019-00314
`Page 17 of 187
`
`
`
`1 system, is then derivable from that
`
`2 information; right?
`
`3 A Essentially, the last sentence
`
`4 of the paragraph says: Additional
`
`5 information, such as centroid,
`
`6 orientation, and bounding box, is derived
`
`7 from this information.
`
`8 Q And that additional information
`
`9 is not actually stored in connection with
`
`10 the object in Kellogg; right?
`
`11 A I believe that's correct, that
`
`12 that's derivable. But whether I --
`
`13 whether that's then stored anywhere or
`
`14 not, I can't say right now. I'd have to
`
`15 look through Kellogg.
`
`16 Q Okay. But you agree that this
`
`17 section that we're looking at -- and feel
`
`18 free to look through it -- does not
`
`19 describe storing that information, but
`
`20 only that it can be derived from what is
`
`21 stored; correct?
`
`22 MR. MARTINELLI: Objection,
`
`23 form.
`
`24 A This doesn't say it's not
`
`25 stored. I believe, to the extent that
`
`AVIGILON EX. 2018
`IPR2019-00314
`Page 18 of 187
`
`
`
`1 these are attributes, Kellogg does
`
`2 disclose storing of attributes. Your
`
`3 question was in this particular section,
`
`4 and I don't see in this particular
`
`5 paragraph that you pointed out that it
`
`6 really goes into storage at all.
`
`7 Q And are you aware of any other
`
`8 part of Kellogg that describes storing
`
`9 those attributes, such as length, surface
`
`10 area, volume, centroid, orientation,
`
`11 bounding box or the like?
`
`12 A I'd have to go through and see
`
`13 where you would talk about storing
`
`14 attributes. I can do that, if you'd
`
`15 like.
`
`16 Q Sure. Why don't you take a
`
`17 look.
`
`18 A On page 50, section 3.6, object
`
`19 storage, through page 52, I believe, it
`
`20 talks about the -- the storage.
`
`21 Q Okay. Can you point me to a
`
`22 passage?
`
`23 A Well, let me look through and
`
`24 pick out a few things, then. So in the
`
`25 preamble to that section it says: The
`
`AVIGILON EX. 2018
`IPR2019-00314
`Page 19 of 187
`
`
`
`1 object-oriented database on which the
`
`2 visual memory builds provides basic
`
`3 support for object storage and retrieval.
`
`4 It says that the visual memory assigns
`
`5 each object and object identifier, and a
`
`6 given object can have multiple versions.
`
`7 It says, for example, a
`
`8 security system could track a person
`
`9 walking down a hall and store a new
`
`10 version describing that person's location
`
`11 every tenth of a second. Then it says,
`
`12 by maintaining all of an object's
`
`13 versions the visual memory can answer
`
`14 questions about the object's history. So
`
`15 that would be storing all of these states
`
`16 of the object every certain amount of
`
`17 time, and then using that later, in a
`
`18 query, to determine the history of the
`
`19 object.
`
`20 And the section goes on for a
`
`21 few pages, here, but there's several
`
`22 indications here of storage, then section
`
`23 3.6.2 specifically describes the storage
`
`24 mechanism, before that it talks about
`
`25 objects appearing and disappearing, and
`
`AVIGILON EX. 2018
`IPR2019-00314
`Page 20 of 187
`
`
`
`1 that that's appropriately stored, that
`
`2 information.
`
`3 So in this -- this goes into
`
`4 details of -- of the storage concept that
`
`5 Kellogg has, in the second paragraph it
`
`6 talks about in certain cases the database
`
`7 should store one base version of the
`
`8 object and then indicate differences for
`
`9 each new version. That's in the case
`
`10 where the object changes very little from
`
`11 moment to moment.
`
`12 Q Right. So none of that,
`
`13 though, describes storing the attributes
`
`14 that we were just talking about, which
`
`15 are, for example, centroid, orientation,
`
`16 length, width, height, area, that sort of
`
`17 thing; right?
`
`18 A Let me go back, I think I
`
`19 mentioned this a moment ago, but it says:
`
`20 Each object can have multiple versions,
`
`21 for example, a security system could
`
`22 track a person walking. So a person
`
`23 could be an attribute. A person walking,
`
`24 that would be another attribute.
`
`25 And it's storing all of this
`
`AVIGILON EX. 2018
`IPR2019-00314
`Page 21 of 187
`
`
`
`1 information, as I just explained a few
`
`2 moments ago. So at least in this section
`
`3 it does indicate that attributes are
`
`4 stored. And we can go through the rest
`
`5 of the Kellogg reference and perhaps find
`
`6 other indications of that.
`
`7 Q Okay. Why don't we turn back
`
`8 to page 24, because I just want to ask
`
`9 about the specific attributes that I was
`
`10 asking about, which, if you go to the end
`
`11 of page 24, and onto 25 it says:
`
`12 Concrete spatial objects can provide
`
`13 additional relevant information; for
`
`14 example, a cube could have functions
`
`15 returning the length of its side, its
`
`16 surface area, and its volume; right?
`
`17 A Yes.
`
`18 Q So as to those variables, what
`
`19 Kellogg teaches is that that information
`
`20 is only obtainable through function calls
`
`21 as defined in the object class; right?
`
`22 MR. MARTINELLI: Object to
`
`23 form.
`
`24 A Can you show me -- can you show
`
`25 me where it says that it can only do
`
`AVIGILON EX. 2018
`IPR2019-00314
`Page 22 of 187
`
`
`
`1 that?
`
`2 Q Well, that's -- the part that
`
`3 we just read is where it says it can do
`
`4 that; right?
`
`5 A Can do that, a moment ago you
`
`6 said only.
`
`7 Q Right. And then so I'd asked
`
`8 you to find any other place where there
`
`9 are other ways to obtain that
`
`10 information. And we went to, I believe
`
`11 it was page 50, and none of these
`
`12 attributes were discussed on page 50; do
`
`13 you agree?
`
`14 A On page 50, those were not an
`
`15 exhaustive list. Those were just
`
`16 examples.
`
`17 Q Okay.
`
`18 A And there's more, for example,
`
`19 on page 52. As a historical database,
`
`20 the visual memory keeps track of when
`
`21 events happened -- this is with regard to
`
`22 the time aspect -- it stores with each
`
`23 version of a temporal object information
`
`24 about that version's time. So this is
`
`25 additional discussion about the database
`
`AVIGILON EX. 2018
`IPR2019-00314
`Page 23 of 187
`
`
`
`1 storing various attributes of an object.
`
`2 Q Right. So none of the
`
`3 attributes that we were discussing on
`
`4 pages 24 to 25, though, are discussed on
`
`5 page 52; right?
`
`6 A These examples, on page 24 and
`
`7 25, are not exhaustive. The examples on
`
`8 page 52 are not exhaustive, nor page 50
`
`9 and 51, as we discussed earlier.
`
`10 Q Okay. But you're not able to
`
`11 point me to any part of Kellogg that
`
`12 describes obtaining parameters related to
`
`13 length, surface area, volume or other
`
`14 measurement information, other than
`
`15 what's described on pages 24 to 25, using
`
`16 function calls; right?
`
`17 A And I just went through pages
`
`18 50 to 52, which do talk about these
`
`19 things. They talk about object movement,
`
`20 positions of objects, so that this --
`
`21 these are measurements of time and space.
`
`22 Q Okay. But they're not
`
`23 measurements of what we've been talking
`
`24 about on pages 24 to 25; do you agree
`
`25 with that?
`
`AVIGILON EX. 2018
`IPR2019-00314
`Page 24 of 187
`
`
`
`1 A I can't agree with that,
`
`2 exactly. For one thing, on page 50, it
`
`3 talked about a person walking. So to
`
`4 determine it's a person, there has to be
`
`5 various information measured from the
`
`6 detected object. And the fact that the
`
`7 person is walking would indicate that
`
`8 there's some measurement of position,
`
`9 which is often done with centroid. So
`
`10 the various low-level parameters that you
`
`11 show on page 24 are elements of the --
`
`12 the higher level of a person or of the
`
`13 activity of a person walking.
`
`14 Q Okay. Will you agree that in
`
`15 order to determine those things on page
`
`16 50 it's not necessary that the items on
`
`17 page 24 to 25 are actually stored, they
`
`18 can just be retrieved using a function;
`
`19 right?
`
`20 A I can't say, right now, that
`
`21 these lower-level attributes, such as
`
`22 centroid, orientation, are -- are not
`
`23 stored. They can be attributes as well
`
`24 as a person walking.
`
`25 So it's not as though we're
`
`AVIGILON EX. 2018
`IPR2019-00314
`Page 25 of 187
`
`
`
`1 saying that only this set of attributes
`
`2 can be stored. Kellogg is offering a --
`
`3 a general approach, here. And I think
`
`4 the user can decide whether or not these
`
`5 things will be stored. And I don't think
`
`6 Kellogg is -- is trying to lock the user
`
`7 down into -- into just this list of
`
`8 things to be stored.
`
`9 Q Okay. And again, you agree
`
`10 that those things you pointed out as
`
`11 attributes, such as person and walking,
`
`12 are only accessible by the system through
`
`13 the data structure of the object that
`
`14 they are associated with; correct?
`
`15 A I didn't say that, no.
`
`16 Q Okay. Do you agree with that?
`
`17 A For example, you mentioned
`
`18 person walking.
`
`19 Q Mm-hmm.
`
`20 A So an -- if person is
`
`21 considered an attribute, and walking is
`
`22 considered an attribute of an object,
`
`23 then these could be stored and they would
`
`24 be accessible as -- as attributes.
`
`25 Q Right. My question is: The
`
`AVIGILON EX. 2018
`IPR2019-00314
`Page 26 of 187
`
`
`
`1 attributes in Kellogg are not accessible
`
`2 as stand-alone variables. They're only
`
`3 accessible through the data structure of
`
`4 the object they're associated with by
`
`5 virtue of being an object-oriented model;
`
`6 true?
`
`7 MR. MARTINELLI: Objection,
`
`8 form.
`
`9 A The database provides these
`
`10 attributes which are stored. The
`
`11 attributes are directly accessible
`
`12 through the database. So there always
`
`13 has to be some way, in any kind of a
`
`14 database, to access the item that you
`
`15 want. And it's -- he doesn't say it's
`
`16 only accessible through that. In fact,
`
`17 let me look further, we've just been
`
`18 looking at page 24, and then the few
`
`19 pages that I mentioned on data storage,
`
`20 because that was the question you had
`
`21 earlier.
`
`22 But let's look at a couple of
`
`23 things, for one, I want to get back to
`
`24 the Kellogg. Let me see what's in the
`
`25 report on that. Well, there's part of
`
`AVIGILON EX. 2018
`IPR2019-00314
`Page 27 of 187
`
`
`
`1 the answer on page 22 of Kellogg, it
`
`2 says: The visual memory design extends
`
`3 the database's storage mechanism. It
`
`4 provides a mechanism for object identity
`
`5 and maintains the history for each
`
`6 object. And then, the last sentence in
`
`7 that paragraph, the design lets
`
`8 applications customize the database
`
`9 storage server based on characteristics
`
`10 of the data they typically store.
`
`11 Q Right. So my question is:
`
`12 When the system wants to access a
`
`13 particular attribute it does so only
`
`14 through the data structure of the object;
`
`15 correct?
`
`16 A And -- no, that's not correct.
`
`17 Q Okay. Please point me to where
`
`18 it discusses doing something other than
`
`19 that.
`
`20 A We'll get to it.
`
`21 Q Dr. Grindon, do you have
`
`22 something specific in mind that you're
`
`23 looking for or are you just looking
`
`24 throughout the Kellogg reference?
`
`25 A I am looking for specific
`
`AVIGILON EX. 2018
`IPR2019-00314
`Page 28 of 187
`
`
`
`1 things.
`
`2 Q Okay. What are you looking
`
`3 for?
`
`4 A In the implementation Kellogg
`
`5 shows that he creates a bucket of
`
`6 attributes. I was looking for that
`
`7 particular page. I have it referenced in
`
`8 my declaration.
`
`9 Q Okay. Are you talking about
`
`10 the bucket index?
`
`11 A The bucket index, yeah. I have
`
`12 that section here. I'm looking in my
`
`13 declaration, right now, for specific
`
`14 reference to that.
`
`15 Q I might be able to short
`
`16 circuit this. If you go to page 83 of
`
`17 Kellogg, it references the bucket index.
`
`18 Are you there?
`
`19 A I'm on page 83.
`
`20 Q If you go down to the last
`
`21 paragraph, it says: In addition to the
`
`22 indices described above, each test also
`
`23 includes a bucket index. A bucket index
`
`24 simply maintains a list of all the
`
`25 objects stored in the visual memory; do
`
`AVIGILON EX. 2018
`IPR2019-00314
`Page 29 of 187
`
`
`
`1 you see that?
`
`2 A Yes.
`
`3 Q Is that what you're referring
`
`4 to?
`
`5 A That's one of the places, yes.
`
`6 Q Now, the bucket index does not
`
`7 maintain a list of attributes, it
`
`8 maintains a list of objects; correct?
`
`9 Do you understand my question,
`
`10 Dr. Grindon?
`
`11 A Yes, I'm just reading this.
`
`12 Q Do you agree with my question:
`
`13 A bucket index maintains a list of
`
`14 objects, not a list of attributes;
`
`15 correct?
`
`16 A I'm -- I'm looking at this.
`
`17 Q Well, the paragraph we just
`
`18 read, at the bottom of 83, literally
`
`19 says: A bucket index simply maintains a
`
`20 list of all the objects stored in the
`
`21 visual memory; right?
`
`22 Can you answer my question?
`
`23 A Your question was limited to
`
`24 that sentence on page 83, and yes, you
`
`25 read it correctly.
`
`AVIGILON EX. 2018
`IPR2019-00314
`Page 30 of 187
`
`
`
`1 Q Okay. You've been looking at
`
`2 this section for awhile, now; is there
`
`3 anything that contradicts that statement?
`
`4 A That's what I was looking for.
`
`5 So on page 75, for example, it
`
`6 says: Temporal indices, section 4.3.3,
`
`7 the prototype temporal indices keep track
`
`8 of the valid times of version -- of
`
`9 object versions. So this index keeps
`
`10 track of the time, which would be an
`
`11 attribute.
`
`12 Q Okay. Can you answer my
`
`13 question, please? When the system wants
`
`14 to access a particular attribute in
`
`15 Kellogg, it does so only through the data
`
`16 structure of the object; correct?
`
`17 A And that appears not to be
`
`18 right. Then, on page 72, under spatial
`
`19 indices, the prototype spatial indices
`
`20 store information about the centroids of
`
`21 objects stored in the visual memory. So
`
`22 these are attributes. In fact, it's one
`
`23 that you had mentioned earlier.
`
`24 Q Where are you looking?
`
`25 A Again, on page 72, section
`
`AVIGILON EX. 2018
`IPR2019-00314
`Page 31 of 187
`
`
`
`1 4.3.2, spatial indices, i-n-d-i-c-e-s.
`
`2 Q Mm-hmm. And how does that
`
`3 contradict my question or how does that
`
`4 -- why does that cause you to disagree
`
`5 with my question?
`
`6 A This spatial index stores the
`
`7 information about the centroids of
`
`8 objects.
`
`9 Q Okay.
`
`10 A So it does not say that you
`
`11 have to go through the object. It says
`
`12 that you've got the information about the
`
`13 centroid. And, of course, the attributes
`
`14 are associated with objects.
`
`15 Q And to access the information
`
`16 about the centroid you would then have to
`
`17 go through the index; correct?
`
`18 A When you say: Go through the
`
`19 index, the index would provide that
`
`20 information.
`
`21 Q Right. The centroid of the
`
`22 object does not exist as a stand-alone
`
`23 variable that the database can access
`
`24 without using the index; right?
`
`25 A So I don't see anywhere in the
`
`AVIGILON EX. 2018
`IPR2019-00314
`Page 32 of 187
`
`
`
`1 claims, and again, we should put the
`
`2 patent in front of us, where it says that
`
`3 you have to do -- you have to access it
`
`4 the way you said.
`
`5 Q Okay. Well, you may disagree
`
`6 that it's relevant, but my question is
`
`7 just about Kellogg. The centroid of the
`
`8 object does not exist as a standalone
`
`9 variable that the database can access
`
`10 without using the index or the object;
`
`11 right?
`
`12 A Without using the --
`
`13 MR. MARTINELLI: Objection
`
`14 to form.
`
`15 A Without using the index? I
`
`16 don't think there's any requirement that
`
`17 you can't use an index to find --
`
`18 Q Okay.
`
`19 A -- the attribute.
`
`20 Q I'm just asking you a question
`
`21 about how Kellogg works. I'm not talking
`
`22 about the claims, okay?
`
`23 So do you agree that the
`
`24 centroid of the object does not exist as
`
`25 a standalone variable that the database
`
`AVIGILON EX. 2018
`IPR2019-00314
`Page 33 of 187
`
`
`
`1 can access without using the object or
`
`2 the index?
`
`3 MR. MARTINELLI: Objection,
`
`4 form.
`
`5 A I don't agree.
`
`6 Q Why not?
`
`7 A Well, first, we're talking
`
`8 about just a particular implementation,
`
`9 and it specifically says that the index
`
`10 stores the centroids. So you have to --
`
`11 if they're stored somewhere you have to
`
`12 be able to go get them where they're
`
`13 stored. And whether you call that an
`
`14 index or something else I don't think is
`
`15 a matter of the claims of the patent.
`
`16 Q Again, I'm not asking about the
`
`17 claims of the patent. I'm asking about
`
`18 the functionality of Kellogg.
`
`19 In this example you pointed me
`
`20 to, at 4.3.2, if the system wants to
`
`21 access the attribute centroid of an
`
`22 object it must either do so through the
`
`23 index or through the object; correct?
`
`24 A It can go to, in this case, the
`
`25 index, and