`
`BEFORE THE PATENT TRIAL AND APPEAL BOARD
`
`META PLATFORMS, INC., )
`
`Petitioner, )
`
`)
`
`)
`
`VS. ) CASE NO.
`
`) IPR2023-00059
`
`ANGEL TECHNOLOGIES GROUP, LLC, )
`
`Patent Owner. ) Volume I
`
`)
`
`___________________________________) Pages 1 through 209
`
`REMOTE DEPOSITION OF DR. ELI SABER
`
`OCTOBER 20, 2023
`
`--ooOoo--
`
`JOB NO.: 6131895
`
`REPORTED BY: MICHELLE MEDEL SABADO, RPR, CRR, CSR #7423
`
`Veritext Legal Solutions
`Calendar-CA@veritext.com 866-299-5127
`
`Page 1
`
`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
`
`Meta Platforms, Inc.
`Meta Platforms, Inc. v. Angel Technologies Group LLC
` IPR2023-0057
`Exhibit 1040 - Page 1 of 283
`
`
`
` REMOTE DEPOSITION OF DR. ELI SABER, TAKEN ON
`BEHALF OF PETITIONER, COMMENCING AT 9:01 A.M. EDT,
`FRIDAY, OCTOBER 20, 2023, AT WEBSTER, NEW YORK, BEFORE
`MICHELLE MEDEL SABADO, RPR, CRR, CSR #7423.
`
`APPEARANCES OF COUNSEL:
`(ALL PARTIES APPEARING REMOTELY)
`
`FOR PETITIONER:
` ALLEN & OVERY, LLP
` BY: ALAN M. BILLHARZ, ESQ.
` 1101 New York Avenue
` Washington, DC 20005
` 202.683.3862
` alan.billharz@allenovery.com
`
`FOR PATENT OWNER:
` MCKOOL SMITH, P.C.
` BY: KAYLEE HOFFNER, ESQ.
` 600 Travis Street, Suite 7000
` Houston, Texas 77002
` 713.485.7320
` khoffner@mckoolsmith.com
`
`Also Present:
`SCOTT W. HEJNY - MCKOOL SMITH, P.C.
`EVAN WAINWRIGHT - ALLEN & OVERY, LLP
`
`Veritext Legal Solutions
`Calendar-CA@veritext.com 866-299-5127
`
`Page 2
`
`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
`
`Meta Platforms, Inc.
`Meta Platforms, Inc. v. Angel Technologies Group LLC
` IPR2023-0057
`Exhibit 1040 - Page 2 of 283
`
`
`
` I N D E X
`
`WITNESS EXAMINATION BY PAGE
`
`DR. ELI SABER
`
` (BY MR. BILLHARZ) 4
`
` E X H I B I T S
`
`NUMBER PAGE DESCRIPTION
`
`EXHIBIT A 61 RIT BULLETIN 1999-00
`
`EXHIBIT B 82 RIT BULLETIN 2003
`
`EXHIBIT C 94 DOCUMENT 39
`
`EXHIBIT D 104 DOCUMENT 36
`
`Veritext Legal Solutions
`Calendar-CA@veritext.com 866-299-5127
`
`Page 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
`
`25
`
`Meta Platforms, Inc.
`Meta Platforms, Inc. v. Angel Technologies Group LLC
` IPR2023-0057
`Exhibit 1040 - Page 3 of 283
`
`
`
` WEBSTER, NEW YORK
`
` Friday, October 20, 2023; 9:01 a.m. Edt
`
` ELI SABER,
`
` having been first duly sworn by the Reporter, was
`
` examined and testified as follows:
`
` EXAMINATION
`
`BY MR. BILLHARZ:
`
` Q Good morning, Dr. Saber.
`
` A Good morning.
`
` Q You understand that you're testifying under
`
`oath today?
`
` A I do.
`
` Q And is there any reason you cannot provide
`
`full, complete and truthful testimony today?
`
` A No, there is not.
`
` Q Okay. Is there anyone in the room with you
`
`besides counsel of record?
`
` A No.
`
` Q Or is there anyone in the room with you at
`
`all?
`
` A No.
`
` Q Now I understand this is not your first
`
`Veritext Legal Solutions
`Calendar-CA@veritext.com 866-299-5127
`
`Page 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
`
`Meta Platforms, Inc.
`Meta Platforms, Inc. v. Angel Technologies Group LLC
` IPR2023-0057
`Exhibit 1040 - Page 4 of 283
`
`
`
`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
`
`deposition; is that correct?
`
` A That is correct.
`
` Q According to your CV, you've been deposed 11
`
`times before; is that right?
`
` A Twelve times.
`
` Q Twelve times.
`
` A Yes. So I probably -- I probably should
`
`update my CV.
`
` Q Understood.
`
` Does that make this lucky number 13?
`
` A I would hope so.
`
` Q Well, given your experience with depositions,
`
`I won't belabor the rules but we'll just cover some
`
`basics. Is that okay?
`
` A I think it's excellent to review the rules.
`
` Q Okay. So first, if I ask a question that you
`
`do not understand, please let me know and I'll try to
`
`correct or clarify the question. If you answer, I'll
`
`presume that you understood my question. Is that fair?
`
` A That's fair.
`
` Q Okay. And to help our court reporter, would
`
`you please answer my questions verbally, no head nods or
`
`gestures?
`
` A I will.
`
` Q Okay. And let's also try not to talk over
`
`Veritext Legal Solutions
`Calendar-CA@veritext.com 866-299-5127
`
`Page 5
`
`Meta Platforms, Inc.
`Meta Platforms, Inc. v. Angel Technologies Group LLC
` IPR2023-0057
`Exhibit 1040 - Page 5 of 283
`
`
`
`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
`
`each other. I know it can be a challenge with these
`
`remote depositions, but let's try not to talk over each
`
`other, okay?
`
` A That would be fantastic.
`
` Q All right. And I will aim to take a break
`
`roughly every hour or so. You can request a break at any
`
`time. But if there's a question pending, I'll ask you to
`
`answer the question before taking a break. Is that fair?
`
` A That's fair. Every hour would be fantastic.
`
` Q Great. So do you understand that today's
`
`deposition is focused on the '432, the '291 and the '275
`
`patents?
`
` A Yes, I do.
`
` Q And when I say '432, '291 and '275 patents,
`
`do you understand what I'm referring to?
`
` A Yes, I do.
`
` Q I'll refer to the corresponding inter partes
`
`review proceedings as the '432 IPR and '291 IPR and the
`
`'275 IPR, okay?
`
` A That would be fine.
`
` Q Now there are four prior art references at
`
`issue that I expect we'll be discussing today. They're
`
`named Sharpe, Eintracht, Carey and Fotofile. If I use
`
`those names, will you understand what I'm referring to?
`
` A Yes, I will.
`
`Veritext Legal Solutions
`Calendar-CA@veritext.com 866-299-5127
`
`Page 6
`
`Meta Platforms, Inc.
`Meta Platforms, Inc. v. Angel Technologies Group LLC
` IPR2023-0057
`Exhibit 1040 - Page 6 of 283
`
`
`
`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
`
` Q Okay. Now you understand that in each of the
`
`'432, '291 and the '275 IPR, Petitioner and Dr. Bederson
`
`have presented multiple obvious grounds based on the
`
`prior art references; is that right?
`
` A That is correct.
`
` Q Okay. So in the '432 IPR, for example,
`
`Petitioner and Dr. Bederson assert that claims one,
`
`three, six, seven and eight of the '432 patent would have
`
`been obvious over Sharpe alone or in view of the
`
`knowledge of a person of ordinary skill in the art. Is
`
`that correct?
`
` A Well, I -- I can obviously check my
`
`declaration but I assume that is correct.
`
` Q Okay. And in the '432 IPR, Petitioner and
`
`Dr. Bederson also assert that claims one through eight of
`
`the '432 patent would have been obvious over Sharpe in
`
`view of Eintracht; is that correct?
`
` A I assume that's correct.
`
` Q Okay. And you understand that obviousness
`
`grounds are identified in the introduction section of
`
`each of Dr. Bederson's declarations in each IPR; is that
`
`right?
`
` A That's correct.
`
` Q Okay. And you also understand that
`
`obviousness grounds are also identified in the
`
`Veritext Legal Solutions
`Calendar-CA@veritext.com 866-299-5127
`
`Page 7
`
`Meta Platforms, Inc.
`Meta Platforms, Inc. v. Angel Technologies Group LLC
` IPR2023-0057
`Exhibit 1040 - Page 7 of 283
`
`
`
`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
`
`identification of challenges section of each IPR
`
`petition?
`
` A That's correct.
`
` Q Now I take it you disagree with the
`
`obviousness grounds presented by Petitioner and
`
`Dr. Bederson?
`
` A I do.
`
` Q Okay. And you have submitted declarations in
`
`each of the '432, '291 and '275 IPR's documenting your
`
`opinions, correct?
`
` A That is correct.
`
` Q So I'll refer to your declarations as the
`
`'432 declaration, the '291 declaration and the '275
`
`declaration, okay?
`
` A That's correct. That's good.
`
` Q And those declarations for the record are
`
`Exhibit 2021 in each proceeding; is that correct?
`
` A I assume so. That would be correct, yeah.
`
` Q And if I refer to them collectively as your
`
`declarations, which I may do from time to time, you'll
`
`understand what I'm referring to?
`
` A Yes, I will.
`
` Q Okay. Did you write your declarations
`
`yourself?
`
` A Yes, I did. I -- I always -- I have a
`
`Veritext Legal Solutions
`Calendar-CA@veritext.com 866-299-5127
`
`Page 8
`
`Meta Platforms, Inc.
`Meta Platforms, Inc. v. Angel Technologies Group LLC
` IPR2023-0057
`Exhibit 1040 - Page 8 of 283
`
`
`
`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
`
`process that I use every single time and in -- in the
`
`many IPR's that I've used in that process that's worked
`
`successfully for me.
`
` Q Would you please describe that process.
`
` A Yes. So I obviously review the materials
`
`very carefully, the -- the patents in question, the prior
`
`art, the references. I have thorough discussions with
`
`counsel about the details of them. I then proceed to --
`
`to put together an initial draft of my opinions and then
`
`we go back and forth on that particular draft until the
`
`declarations are finalized and then I review the
`
`declarations very, very carefully, finalize everything,
`
`touch up any details that are needed and then sign and
`
`finalize.
`
` Q So Doctor, you said you stand by your
`
`opinions in your '432, '291 and '275 declarations,
`
`correct?
`
` A I do.
`
` Q Do you have any corrections to your
`
`declarations at this time?
`
` A No, I do not. I saw a couple of typos here
`
`and there but I'll point them out when they show up.
`
` Q That would be great. Thank you.
`
` There is some overlap across your
`
`declarations. When I ask you about your opinion
`
`Veritext Legal Solutions
`Calendar-CA@veritext.com 866-299-5127
`
`Page 9
`
`Meta Platforms, Inc.
`Meta Platforms, Inc. v. Angel Technologies Group LLC
` IPR2023-0057
`Exhibit 1040 - Page 9 of 283
`
`
`
`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
`
`regarding a particular position, I may only reference one
`
`section of your declarations but I am asking about your
`
`opinion overall across all three declarations, is that
`
`fair?
`
` A That would be fair but if I'm not clear, I
`
`will clarify.
`
` Q Yes. Absolutely, please. If you -- if your
`
`answer would be different for one of the three patents at
`
`issue today, please let me know.
`
` A Yes.
`
` Q And -- and the other side of that coin is if
`
`my question is specific to one of the patents, I'll
`
`indicate that with the question. Is that fair?
`
` A That's fair.
`
` Q Okay. What did you do to prepare for today's
`
`deposition?
`
` A Well, I reviewed the materials, you know, the
`
`petitions, the declarations, the prior art, the patents
`
`obviously in question and I've had discussions with
`
`counsel.
`
` Q Did you review Dr. Bederson's deposition
`
`transcript in preparation for today's deposition?
`
` A I have.
`
` Q Did you review Dr. Bederson's declarations in
`
`preparing for today's deposition?
`
`Veritext Legal Solutions
`Calendar-CA@veritext.com 866-299-5127
`
`Page 10
`
`Meta Platforms, Inc.
`Meta Platforms, Inc. v. Angel Technologies Group LLC
` IPR2023-0057
`Exhibit 1040 - Page 10 of 283
`
`
`
`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
`
` A I have.
`
` Q And when you reviewed the prior art
`
`references, did you review them in their entirety?
`
` A Yes, I did.
`
` Q And when you reviewed the patents at issue
`
`today, did you review them in their entirety?
`
` A Yes, I did.
`
` Q And roughly, how many hours did you spend
`
`preparing for today's deposition?
`
` A I don't remember off the top of my head, but
`
`in this case, I probably have somewhere between 60 and 70
`
`hours at this point.
`
` Q And are you being compensated at an hourly
`
`rate?
`
` A I am.
`
` Q Is your compensation in any way tied to the
`
`opinions that you're offering in these IPR's?
`
` A No, it's not.
`
` Q Is your compensation tied in any way to the
`
`outcome of these IPR's?
`
` A No, it is not.
`
` Q Do you have any financial interest in the
`
`patent owner Angel Technologies Group?
`
` A No, I do not.
`
` Q Do you have any relationship with the named
`
`Veritext Legal Solutions
`Calendar-CA@veritext.com 866-299-5127
`
`Page 11
`
`Meta Platforms, Inc.
`Meta Platforms, Inc. v. Angel Technologies Group LLC
` IPR2023-0057
`Exhibit 1040 - Page 11 of 283
`
`
`
`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
`
`inventor on the '432, '291 and '275 patents? That's Mark
`
`Frigon?
`
` A No, I do not.
`
` Q Do you have any relationship with Chris
`
`Malone who submitted a declaration in these proceedings?
`
` A No.
`
` Q Do you have any relationship with Lisa Larson
`
`who also submitted a declaration in these proceedings?
`
` A No, I do not.
`
` Q So Dr. Saber, would you please describe your
`
`educational background.
`
` A Yes. I received a associate in science in
`
`engineering science from the Mohawk Valley Community
`
`College in 1986. I received a bachelors degree in --
`
`bachelor of science degree in electrical and computer
`
`engineering from the University of Buffalo in 1988. I
`
`received a masters of science in electrical and computer
`
`engineering from the University of Rochester in 1992.
`
`And I received a Ph.D in electrical and computer
`
`engineering from the University of Rochester in 1996.
`
` Q Okay. And did you specialize in any
`
`particular area in your undergraduate studies?
`
` A My undergraduate studies was kind of a
`
`general but with some specialization. I -- I spent a lot
`
`of time on the hardware type of things, control systems,
`
`Veritext Legal Solutions
`Calendar-CA@veritext.com 866-299-5127
`
`Page 12
`
`Meta Platforms, Inc.
`Meta Platforms, Inc. v. Angel Technologies Group LLC
` IPR2023-0057
`Exhibit 1040 - Page 12 of 283
`
`
`
`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
`
`microprocessors. I've done some signal processing as
`
`well and, you know, circuitry, just you know, general
`
`electrical engineering. I've taken several communication
`
`courses at the undergraduate and the graduate level at
`
`that time.
`
` Q And did you work with databases at all at an
`
`undergrad?
`
` A I don't remember off the top of my head if I
`
`did any projects related to databases as an undergrad.
`
`It's been -- it's been a long time. It's been a long
`
`time.
`
` Q Yes. I was going to ask you about websites
`
`but 1998 might be a little too early.
`
` A 1988.
`
` Q 1988, apologies.
`
` A Yes.
`
` Q So let me ask you about in your masters
`
`studies, did you have a particular specialization there?
`
` A As I started my masters degree in -- in
`
`electrical engineering, I started to specialize more in
`
`the signal image video processing computer vision area
`
`and also the communication systems, you know, that
`
`working area. So my specialities, you know, started to
`
`move away from the hardware in a sense and
`
`instrumentation, more towards that type of area.
`
`Veritext Legal Solutions
`Calendar-CA@veritext.com 866-299-5127
`
`Page 13
`
`Meta Platforms, Inc.
`Meta Platforms, Inc. v. Angel Technologies Group LLC
` IPR2023-0057
`Exhibit 1040 - Page 13 of 283
`
`
`
`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
`
` Q So what drove you --
`
` A And controls.
`
` Q What motivated your movement towards images
`
`and imaging technologies?
`
` A Well, I -- as I started taking graduate
`
`courses in the different areas, I started to discover
`
`that I have more of a, you know, interest in that area.
`
`I like, obviously, the mathematics. Quite a bit of the
`
`area, the signal processing and communications and
`
`networking area and controls come with a heavier dose of
`
`mathematics, if you would, versus instrumentation and
`
`hardware type areas which I find myself to be, you know,
`
`really enjoying so I started to gravitate more towards
`
`those areas.
`
` Q And you mentioned networking. Can you
`
`elaborate a little bit on the type of networking you were
`
`doing during your masters studies.
`
` A Well, I took courses in communication
`
`systems, you know, detaching and estimation theory, you
`
`know, and when you take a course in communication systems
`
`it includes, you know, networking type of information
`
`like queuing and, you know, that type of thing. I took
`
`course in controls as an undergrad in that area.
`
` Q Okay.
`
` A I think I still have the book in my office.
`
`Veritext Legal Solutions
`Calendar-CA@veritext.com 866-299-5127
`
`Page 14
`
`Meta Platforms, Inc.
`Meta Platforms, Inc. v. Angel Technologies Group LLC
` IPR2023-0057
`Exhibit 1040 - Page 14 of 283
`
`
`
`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
`
` Q That's great.
`
` And during the master studies, do you recall
`
`any work with databases?
`
` A I started working with databases because I
`
`started working on digital libraries in my masters and
`
`then I continued that obviously into my Ph.D., but as you
`
`can appreciate, when you're taking masters courses at the
`
`University of Rochester, at least every course they
`
`require to you do some type of project. So the professor
`
`would like you to do -- essentially create -- you know,
`
`provide a proposal of a project that you'd like to
`
`undertake and then undertake the project and then do a
`
`presentation to your colleagues in the classroom about
`
`the project.
`
` So obviously, as I started undertaking
`
`courses more in the communications and, you know,
`
`networking area and the image processing and video
`
`processing area and the controls area I started to create
`
`projects in those areas, with the intent of obviously
`
`learning more about the areas that interested me more and
`
`more so I can build that into essentially a dissertation
`
`for Ph.D.
`
` And I also started working with my advisor at
`
`the time, Dr. Murat Tekalp, who was essentially a full
`
`professor at the time and really the head -- the head of
`
`Veritext Legal Solutions
`Calendar-CA@veritext.com 866-299-5127
`
`Page 15
`
`Meta Platforms, Inc.
`Meta Platforms, Inc. v. Angel Technologies Group LLC
` IPR2023-0057
`Exhibit 1040 - Page 15 of 283
`
`
`
`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
`
`the signal image video processing and computer vision
`
`area in the department. So Dr. Tekalp started to
`
`motivate me more and more into that area because
`
`obviously, Dr. Tekalp was looking for students to
`
`continue the work that he -- he's been undertaking.
`
` Q You mentioned a digital libraries project.
`
`Was that your masters project?
`
` A Well, I started it in MS on a smaller scale
`
`and I basically grew it extensively in my Ph.D. work, but
`
`the MS work started the groundwork for it in a sense
`
`because obviously, I wanted to take my MS time to learn
`
`more about the area as much as I can. You know, the
`
`masters essentially provides the foundation in terms of
`
`course work. So ideally speaking, you would like to
`
`place your course work as much as possible in the area
`
`where you think you're going to undertake the Ph.D work
`
`because the course work provides all the foundation, the
`
`mathematics, the up-to-date -- you know, you want to be
`
`up-to-date in terms of, you know, algorithms and all the
`
`things that are needed. So when you start the Ph.D work,
`
`you know, you're at least -- you're working at the state
`
`of the art.
`
` You now can take a look at the state of the
`
`art, see where the issues are, where the challenges are,
`
`what's missing and then undertake a Ph.D study to try to
`
`Veritext Legal Solutions
`Calendar-CA@veritext.com 866-299-5127
`
`Page 16
`
`Meta Platforms, Inc.
`Meta Platforms, Inc. v. Angel Technologies Group LLC
` IPR2023-0057
`Exhibit 1040 - Page 16 of 283
`
`
`
`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
`
`advance the state of the art which will hopefully result
`
`in original publications that you would, you know,
`
`publish in conferences and archive in journals. So
`
`that's the ideal model. So if you can use your masters
`
`to build up the knowledge, then you don't have to spend
`
`your time in the Ph.D building up the knowledge.
`
` Q In terms of a foundation in your masters
`
`degree, is there a particular course that is the
`
`foundation for image processing work?
`
` A Oh, absolutely. I mean, I -- I -- at U of R
`
`I remember my course work correctly and again, it's been
`
`quite awhile.
`
` Q Understood.
`
` A So I took digital signal processing at the
`
`graduate level which is essentially the one dimensional
`
`version. Then I took digital image processing at the
`
`graduate level which provides the two dimensional
`
`version, then I took digital video processing at the
`
`graduate level which now takes the imagery into a
`
`temporal domain so you have, you know, images as a
`
`function of time in a sense and it brings in a whole slew
`
`of algorithms and things that have to do with the
`
`temporal domain.
`
` I took pattern recognition at the graduate
`
`level as well which brings in a whole host of the
`
`Veritext Legal Solutions
`Calendar-CA@veritext.com 866-299-5127
`
`Page 17
`
`Meta Platforms, Inc.
`Meta Platforms, Inc. v. Angel Technologies Group LLC
` IPR2023-0057
`Exhibit 1040 - Page 17 of 283
`
`
`
`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
`
`mathematical algorithms and things that, you know, where
`
`pattern recognition essentially cuts through a lot of
`
`areas because pattern recognition can be face
`
`recognition, it can be object recognition, it can be a
`
`lot of things. I took computer vision as well at the
`
`graduate level. All of my courses were at the graduate,
`
`obviously, when I was in the MS program.
`
` So I took computer vision as well. I took
`
`communication theory at the graduate level which
`
`included, you know, a section on networking, obviously.
`
`I took detection and estimation theory at the graduate
`
`level. I took more course work actually than I'm
`
`supposed to take. MS is supposed to be 30 credits which
`
`will amount to ten courses. I think I ended up with
`
`maybe 13 or 14 courses so I took a little bit more course
`
`work than are needed because I wanted, you know, to make
`
`sure I knew the state of the art and I had an
`
`understanding as wide as possible so when I dove into the
`
`Ph.D work, you know, I was well prepared.
`
` Q I'd like to take a step back to that digital
`
`libraries project --
`
` A Okay.
`
` Q -- that you mentioned earlier. You said you
`
`started it during your masters studies. Would you
`
`describe what you were doing with that, maybe describe
`
`Veritext Legal Solutions
`Calendar-CA@veritext.com 866-299-5127
`
`Page 18
`
`Meta Platforms, Inc.
`Meta Platforms, Inc. v. Angel Technologies Group LLC
` IPR2023-0057
`Exhibit 1040 - Page 18 of 283
`
`
`
`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
`
`the project a little bit?
`
` A Sure. So the idea behind the project
`
`initially started with it -- if you were to go let's say
`
`on a vacation, Disney World or a ski trip or you went on
`
`a vacation and you had your digital camera with you and
`
`you were just shooting pictures and let's say you had
`
`10,000 pictures and you came back home and you plugged
`
`the camera into your computer and you downloaded those
`
`10,000 pictures into a folder. Now you want to search
`
`them. You want to be able to look at different things in
`
`these images, you know -- find you know, the Ferrari that
`
`you were standing next to at Disney World or find that
`
`magical moment where you were standing next to Mickey
`
`Mouse or whatever.
`
` You want to find imagery. You want to locate
`
`objects and things in these imagery. How would you do
`
`the search, you know? Essentially, you know, the -- the
`
`slogan I used to use when I started presenting some of
`
`this work is we used to call our system Voogle and we
`
`essentially got the name because I took the word Google,
`
`I took "G" off of it and put a "V" in front of it and
`
`kind of called it visual Google. So the idea was images
`
`are to Voogle as text is to Google.
`
` That's the slogan I used to try to sell as I
`
`was presenting at conferences at various -- at various,
`
`Veritext Legal Solutions
`Calendar-CA@veritext.com 866-299-5127
`
`Page 19
`
`Meta Platforms, Inc.
`Meta Platforms, Inc. v. Angel Technologies Group LLC
` IPR2023-0057
`Exhibit 1040 - Page 19 of 283
`
`
`
`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
`
`you know, settings around the world, international. That
`
`is where the idea initially started is, how would you
`
`access the imagery? How would you present -- how would
`
`you present your query to the images? Like how would you
`
`say I'm looking for this Ferrari in -- in these 10,000
`
`images or a hundred thousand, however many you shot?
`
` Or it could -- it doesn't have to be one
`
`trip, right? You might have gone on five, six, eight
`
`trips this year and had folder one for trip one, folder
`
`two for trip two, folder three for trip three. How would
`
`you access these different folders and search for a
`
`variety of different things? I saw a Ferrari in Disney
`
`World and I saw one in Utah on a ski trip and I want both
`
`pictures up when I show you what a Ferrari looks like,
`
`that type of ideas.
`
` Q Was the digital -- or did you implement the
`
`digital libraries project in code?
`
` A Yes. Oh, yes. I spent a lot of time writing
`
`C and a little bit of C++ but a lot of C at the time.
`
`Some C++ for the right objects but a lot of C. C was
`
`essentially a really good engineering language because it
`
`has all the things that you need to -- to do, you know,
`
`engineering work unlike, for instance, Pascal or some of
`
`the other languages. So C is very, very well suited for
`
`it. The only sometimes challenge is you have to write
`
`Veritext Legal Solutions
`Calendar-CA@veritext.com 866-299-5127
`
`Page 20
`
`Meta Platforms, Inc.
`Meta Platforms, Inc. v. Angel Technologies Group LLC
` IPR2023-0057
`Exhibit 1040 - Page 20 of 283
`
`
`
`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
`
`also all the underlying framework. Like if you were to
`
`add complex numbers, you have to write that library. If
`
`you want to -- or get somebody else's library but then
`
`you don't know how well -- maybe there's a bug here or
`
`there. You don't know.
`
` So you have to either write the underlying
`
`framework or, you know, get it from somebody else. Like
`
`if you want to Fourier transform algorithms or inverse
`
`Fourier transform, you have to write those or acquire
`
`them so it takes a bit to build up the background a bit
`
`and I -- and I wrote, you know, a lot of background stuff
`
`at the lab because the lab had already some. Like
`
`Fourier transforms and inverse Fourier transforms were
`
`there, so I could use them directly but a lot of the
`
`materials that I needed like, you know, principal
`
`component analysis or some other things I actually wrote
`
`myself and I had a pretty extensive library at the time
`
`of software by the time I graduated.
`
` Q Where are the photos stored in the digital
`
`libraries project?
`
` A I think we had a server. It's been awhile,
`
`you know. I think we had a server at -- at the --
`
`because, you know, there were a lot of photos so you
`
`can't store them. And the computers at the time didn't
`
`have terabytes of data storage, so you know, we were
`
`Veritext Legal Solutions
`Calendar-CA@veritext.com 866-299-5127
`
`Page 21
`
`Meta Platforms, Inc.
`Meta Platforms, Inc. v. Angel Technologies Group LLC
` IPR2023-0057
`Exhibit 1040 - Page 21 of 283
`
`
`
`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
`
`working on SPARC 10's, unit SPARC 10 stations at the
`
`time. You know, the PC was a little on the early side if
`
`you would. So we had I think -- I think the lab had a
`
`server where we stored all the stuff but I don't remember
`
`exactly off the top of my head.
`
` Q Was -- was a digital libraries project based
`
`on the database?
`
` A Well, the images essentially were, you know,
`
`essentially placed into folders and as we started doing
`
`the annotations and that type of thing, you have to
`
`somehow get some relationship because you don't want to
`
`every time have to do -- you don't want to run your
`
`algorithm every time and have to go back in and have to
`
`annotate all of them all over again because that takes a
`
`lot of time, especially at that timeframe when, you know,
`
`computers weren't as fast as they are today.
`
` So as you've done work on the images, you --
`
`you store that information in some type of a -- you know,
`
`some type of database that you would need to be able to
`
`do the extraction a little quicker than -- than you --
`
`you would have to if you had to run everything all over
`
`again from scratch every single time.
`
` Q So just to make sure I understand, are you
`
`saying that the -- the images were stored in folders but
`
`the annotation information was stored in the database?
`
`Veritext Legal Solutions
`Calendar-CA@veritext.com 866-299-5127
`
`Page 22
`
`Meta Platforms, Inc.
`Meta Platforms, Inc. v. Angel Technologies Group LLC
` IPR2023-0057
`Exhibit 1040 - Page 22 of 283
`
`
`
`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
`
` A Yeah, so the images would be stored in a --
`
`you know, folders because we had many and, you know, in
`
`some type of what we -- a database that kind of contained
`
`all the imagery because they were all raster data in a
`
`sense, you know, color and grayscale. And the
`
`information that you extract from the images, any text --
`
`any -- any textual annotations or any kind of connections
`
`between the images, you would store that into some type
`
`of a database, you know. I wrote some of that code
`
`myself so I can essentially -- I know how to access it
`
`and go in and out on those details.
`
` Q Would you please describe these annotations?
`
`What do they look like?
`
` A Well, some of them were textual annotations.
`
`For instance, let's say you had an image and you will say
`
`Eli's in this image and John Doe is in this image and in
`
`this image, you would provide some textual annotations
`
`that would mark information in that image. If the image
`
`was an outdoor image, you would provide some annotations
`
`that would say this image is an outdoor image or an
`
`indoor image or, you know, I see outdoor scenery or this
`
`is more of a street type imagery versus, you know,
`
`mountain type imagery, that type of thing. So there are
`
`some textual annotations that have to do with the content
`
`of the imagery.
`
`Veritext Legal Solutions
`Calendar-CA@veritext.com 866-299-5127
`
`Page 23
`
`Meta Platforms, Inc.
`Meta Platforms, Inc. v. Angel Technologies Group LLC
` IPR2023-0057
`Exhibit 1040 - Page 23 of 283
`
`
`
`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
`
` But there's also some -- some visual type of
`
`information as well because if you had, for instance, a
`
`car in the image and you want to know where the car is
`
`exactly, you can't put that in text so you have to have
`
`some type of segmentations that would, you know, have
`
`templates that segment the imagery and provide in a sense
`
`false coloring of various objects in the imagery so the
`
`segmentation profiles are also stored with the imagery,
`
`stored, you know, in a sense in a database and related to
`
`the particular image. So every time you segment