IPR2016-00212, Paper No. 41
`March 8, 2017
`
`trials@uspto.gov
`571-272-7822
`
`
`
`RECORD OF ORAL HEARING
`UNITED STATES PATENT AND TRADEMARK OFFICE
`- - - - - -
`BEFORE THE PATENT TRIAL AND APPEAL BOARD
`- - - - - -
`GOOGLE INC.,
`Petitioner,
`v.
`VEDANTI LICENSING LIMITED,
`Patent Owner.
`- - - - - - -
`Case IPR2016-002121
`Patent 7,974,339 B2
`Technology Center 2400
`Oral Hearing Held: Tuesday, February 14, 2017
`
`Before: MICHAEL R. ZECHER, JUSTIN T. ARBES, and
`JOHN A. HUDALLA, Administrative Patent Judges.
`
`The above-entitled matter came on for hearing on Tuesday,
`February 14, 2017, at 10:00 a.m., in Hearing Room A, taken at the U.S.
`Patent and Trademark Office, 600 Dulany Street, Alexandria, Virginia.
`
`
`1 Case IPR2016-00215 has been consolidated with this proceeding.
`
`
`March 8, 2017
`
`trials@uspto.gov
`571-272-7822
`
`
`
`RECORD OF ORAL HEARING
`UNITED STATES PATENT AND TRADEMARK OFFICE
`- - - - - -
`BEFORE THE PATENT TRIAL AND APPEAL BOARD
`- - - - - -
`GOOGLE INC.,
`Petitioner,
`v.
`VEDANTI LICENSING LIMITED,
`Patent Owner.
`- - - - - - -
`Case IPR2016-002121
`Patent 7,974,339 B2
`Technology Center 2400
`Oral Hearing Held: Tuesday, February 14, 2017
`
`Before: MICHAEL R. ZECHER, JUSTIN T. ARBES, and
`JOHN A. HUDALLA, Administrative Patent Judges.
`
`The above-entitled matter came on for hearing on Tuesday,
`February 14, 2017, at 10:00 a.m., in Hearing Room A, taken at the U.S.
`Patent and Trademark Office, 600 Dulany Street, Alexandria, Virginia.
`
`
`1 Case IPR2016-00215 has been consolidated with this proceeding.
`
`
`
`APPEARANCES:
`
`ON BEHALF OF THE PETITIONER:
`
`
`
`
`
`
`
`
`
`
`
`MICHAEL V. MESSINGER, ESQ.
`BRIAN W. LEE, ESQ.
`Sterne, Kessler, Goldstein & Fox PLLC
`1100 New York Avenue, N.W.
`Suite 600
`Washington, D.C. 20005
`202-371-2600
`
`
`
`
`
`
`
`
`ON BEHALF OF THE PATENT OWNER:
`
`
`
`
`
`
`
`
`ROBERT M. ASHER, ESQ.
`BRANDON SCRUGGS, ESQ.
`Sunstein Kann Murphy & Timbers LLP
`125 Summer Street
`Boston, Massachusetts 02110
`617-443-9292
`
`2
`
`
`
`
`
`
`
`
`
`
`
`
`
`P R O C E E D I N G S
`
`Case IPR2016-00212
`Patent 7,974,339 B2
`
`
`
` (10:00 a.m.)
`JUDGE HUDALLA: Please be seated, everyone.
`Good morning. We will hear argument now in IPR2016- 00212,
`pertaining to U.S. Patent Number 7,974,339. Case
`IPR2016-00215 has been consolidated with this proceeding and
`it pertains to the same patent.
`Who do we have appearing for Petitioner today?
`MR. MESSINGER: Michael Messinger, Your
`
`Honor.
`
`Messinger.
`
`JUDGE HUDALLA: Good morning, Mr.
`
`Patent Owner?
`MR. ASHER: Robert Asher.
`JUDGE HUDALLA: Good morning, Mr. Asher.
`Okay. Well, thank you and welcome. And per our
`order dated January 17, 2017, each party will have 50 minutes
`to argue today. And because we have a Motion to Amend
`pending, we're going to have potentially a four-part agenda
`today.
`
`First, we will hear from Petitioner presenting its
`case- in-chief as to the challenged claims. Petitioner can
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`
`
`
`3
`
`
`
`Case IPR2016-00212
`Patent 7,974,339 B2
`
`
`
`reserve rebuttal time. Patent Owner will then respond to
`Petitioner's case and will present its case regarding the Motion
`to Amend, and Patent Owner may reserve rebuttal time as to
`Petitioner's arguments regarding the Motion to Amend.
`Third, we will have Petitioner using the rest of its
`time to respond to Patent Owner's presentation and, fourth,
`Patent Owner may then use the remaining time to respond to
`Petitioner's arguments regarding the Motion to Amend, if any
`are made.
`
`I will remind the parties that Petitioner bears the
`burden of proving any proposition of unpatentability by a
`preponderance of the evidence. And I will also remind the
`parties that we have a court reporter today transcribing the
`proceedings and the record will become public after the
`hearing.
`
`I also remind the parties to please mention the
`slide number if you are going to be discussing it so that we
`will have some clarity in the record as we go forward.
`So with that I guess, Mr. Messinger, if you would
`like to proceed. Would you like to reserve some time today?
`MR. MESSINGER: Yes. 20 minutes, Your Honor.
`JUDGE HUDALLA: Okay.
`
`
`
`4
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`
`
`
`Case IPR2016-00212
`Patent 7,974,339 B2
`
`
`
`MR. MESSINGER: May it please the Board.
`Happy Valentine's Day. Let me first give an overview and then
`I will address the two grounds.
`The '339 patent boils down to two steps, dividing a
`frame of image data into different regions based on level of
`detail and then selecting a pixel from each region for
`transmission. This simple sampling was well known in the
`prior art and set out in Google's petition. Go to slide 2.
`The Board correctly instituted trial on two grounds
`of obviousness for the '339 patent, namely that the challenged
`claims are obvious over Spriggs in view of Golin and the
`second ground Belfor in view of Thyagarajan and further in
`view of Golin. None of the argument or evidence Patent
`Owner has presented since disturbs this result.
`Slide 3. For the Spriggs ground Patent Owner
`simply misreads the teachings of Spriggs as laid out in the
`petition. Spriggs plainly teaches both the analysis and the
`pixel selection as claimed, and Spriggs also teaches the pixel
`data, that is the pixel values, and region data, which are the
`division codes and coordinate addresses for the corners that are
`forming a block in Spriggs.
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`
`
`
`5
`
`
`
`Case IPR2016-00212
`Patent 7,974,339 B2
`
`
`
`And contrary to what the Patent Owner argues, this
`is all laid out in our petition as I will demonstrate when we
`walk through and supported by the testimony, the original
`testimony of Dr. Grindon. For the Belfor ground Patent Owner
`mainly tries to show that there is no motivation to combine the
`references.
`However, as Google showed in its petition, Belfor's
`approach of dividing a frame into uniform size blocks can be
`substituted with Thyagarajan's subdivision of blocks involving
`non-uniform sizes. Belfor even suggests non-uniform size
`blocks, and the Thyagarajan teaching on subdivision, it is
`drawn from the block size assignment as we will get into, and
`that occurs before DCT, Discrete Cosine Transform,
`processing, and what the petition is relying on is that
`subdivision from the block size assignment.
`And then just one point, the combination of
`references applied in the petition also teaches the region data,
`the data receiving system as in claim 1 and the assembling as
`in the method steps of 7 and 10.
`So let me touch on the '339 patent and then I will
`go into the grounds in more detail. So if we look at slide 4,
`fig. 1 shows the data transmission system in one embodiment
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`
`
`
`6
`
`
`
`Case IPR2016-00212
`Patent 7,974,339 B2
`
`
`
`and includes the frame analysis system 106 and the pixel
`selection system 108.
`There is also the data receiving system 104 which
`has the pixel data system and the display generation system.
`And this sort of selection of the pixels is what allows the
`reduced amount of image data to be transmitted without having
`to be compressed.
`If you look at slide 5, the way the '339 patent
`works is it does divide up a region and there are two
`embodiments. One is in fig. 9 where it is sort of uniform
`regions or uniform matrices in particular in this example in fig.
`9. And in fig. 10 there is an embodiment where the matrices
`can have non-uniform sizes.
`So let's just touch on the claims and then get into
`the grounds and show why they are obvious. Claim 1 on slide
`6, it is a system claim, and this is important, Your Honor, in
`that the language that is used to describe the analysis system
`and the language that is used to describe the pixel selection, it
`is cast in fairly broad language that just recites receiving data
`and then generating data. It is sort of the form of these claims.
`So for the analysis system it is receiving frame
`data and there is generating region data. And with the pixel
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`
`
`
`7
`
`
`
`Case IPR2016-00212
`Patent 7,974,339 B2
`
`
`
`selection system it receives region data and generates a set of
`pixel data forming a new set of data for transmission. A pretty
`high level. As we will show, we will demonstrate, the prior art
`teaches exactly that.
`So why don't we look at the first ground, slide 9,
`Spriggs in view of Golin. And actually we can jump to slide
`10 in the interest of time. Spriggs discloses a coding process
`for a transmitter that reduces the image data transmitted, and
`you can see that on fig. 5. That's what we relied on in our
`petition.
`
`There is a dispute with respect to the coding
`process and how that is -- the different aspects of that that read
`on the analysis and the pixel selection, but it is very clear and
`both parties agree that is operating in the processor here.
`JUDGE HUDALLA: Can I interrupt you there?
`We are looking at figure 5. Which portions of figure 5 do you
`contend are the analysis system and which portions do you
`contend are the pixel selection system?
`MR. MESSINGER: Yes, with respect to fig. 5,
`both of those aspects of the coding process are carried out by
`the processor 14. You can see the frame store, which is part of
`the analysis and that receives the frame data, but the aspect of
`
`
`
`8
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`
`
`
`Case IPR2016-00212
`Patent 7,974,339 B2
`
`
`
`the coding process is carried out in the processor 14 and the
`aspect of the pixel selection is carried out by the processor 14.
`As we get into it, there is additional description in
`Spriggs of this sort of process that it goes through that actually
`identifies different aspects of that process correspond to those
`elements. I can step through that. We can see that in the next
`slide, slide 11.
`JUDGE ZECHER: Counselor, can we take you
`back real quick and look at the language of claim 1. It talks
`about analysis on a pixel selection system.
`In light of the specification do you think that those
`two systems are mutually exclusive or can they be one and the
`same?
`
`MR. MESSINGER: I'm sorry, in light of the
`specification?
`JUDGE ZECHER: In light of the specification of
`the '339 patent, do you think that the analysis system and the
`pixel selection system are mutually exclusive systems or can
`they be one and the same?
`MR. MESSINGER: They can -- there is discussion
`in the '339 patent of these systems, like the analysis and the
`
`
`
`9
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`
`
`
`Case IPR2016-00212
`Patent 7,974,339 B2
`
`
`
`pixel selection, basically operating on a computer. There is
`actually a number of different embodiments the '339 describes.
`If you look, for example, at column 3 of the '339
`patent, lines 3 to 12, it even describes that the software system
`can be one or more objects. It even says lines of code. And so
`our position is that each of these systems can be just different
`lines of code, different functions that are carried out by the
`coding process.
`And the overall '339, there is nothing in it that
`would preclude a processor such as processor 14 from reading
`on the '339 implementation.
`JUDGE ZECHER: Thank you.
`JUDGE ARBES: But, counsel, you would agree
`that you have to be able to distinguish one from the other, even
`if they are different operations performed by the same
`executing program, I have to be able to distinguish A from B,
`right?
`
`MR. MESSINGER: Yes, again, referring to that
`general claim language, it has to read on that claim language
`and that is pretty broad claim language. We did recognize in
`our petition and the evidence that there is two elements there.
`
`
`
`10
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`
`
`
`Case IPR2016-00212
`Patent 7,974,339 B2
`
`
`
`There is an analysis system and there is a pixel selection
`system.
`
`Patent Owners then argue that somehow there is
`sort of a separately identifiable kind of additional limitation
`that needs to be read into that. We think that is unwarranted
`given the claim language.
`However, even then our petition and the evidence
`of Spriggs shows that the part of the algorithm in Spriggs
`where there is an interpolation, and then if the block has
`decided that subdivision is needed, it actually generates corner
`addresses and then it puts those corner addresses for the blocks
`that are going to be subdivided onto a stack.
`And that aspect of the coding process reads on the
`analysis system. And then the coding process continues and
`actually talks about how there is a division code 1 that is
`transmitted, and then there is additional corner samples for
`those subdivided regions that are transmitted, and those aspects
`read squarely on the pixel selection which talks about for each
`region.
`
`JUDGE ARBES: So just to make sure I understand
`your position, there is a coding process that's performed by
`processor 14 and certain operations correspond to the analysis
`
`
`
`11
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`
`
`
`Case IPR2016-00212
`Patent 7,974,339 B2
`
`
`
`system and then certain operations correspond to the pixel
`selection system. Is that right?
`MR. MESSINGER: Yes, correct, Your Honor.
`And I could even go into more detail on sort of how we do that
`mapping, but that's exactly how.
`JUDGE ARBES: And is there enough detail in
`Spriggs to say these are the specific operations that correspond
`to A and these are the specific operations that correspond to B?
`MR. MESSINGER: Yes. In fact, it shows up in
`fig. 6, which is in our petition for the pixel selection system.
`We can go -- well, here is the division, on slide 11, here is
`where it is talking about the actual division of the frame data
`into regions of level of detail which meets the claim language
`of the analysis system.
`And that's an interpolation. And if the values have
`a variance amongst the pixels, then it is subdivided. And
`pixel, as you noted in your Institution Decision, Spriggs has
`extensive teaching on sort of how that subdivision happens.
`And then if you look at fig. 6 this shows, further
`shows how the pixel selection system is carried out in a
`different part of the coding process. Once the divisions have
`been made and those corner addresses of a subdivided region
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`
`
`
`12
`
`
`
`Case IPR2016-00212
`Patent 7,974,339 B2
`
`
`
`identified, then a division code is sent. And that's in, like, for
`example, row 1 where it says SE, SF, SG, if the interpolation
`shows that there is going to be a subdivision, then the corner
`addresses of those four new blocks that are going to be
`subdivided -- that are subdivided, those are stored on a stack --
`and I will show you fig. 4 which shows that in more detail if
`you like -- but here what happens is those corner addresses go
`in a stack. That's the analysis system.
`And then the process continues and the division
`code 1 on the first row, for example, is transmitted, and then
`the corner samples for those regions that are subdivided, which
`in this case is SE, SF, SG, SI, SH, those are transmitted. Fig.
`4 actually puts it in flow chart form and lays out the different
`aspects.
`
`JUDGE HUDALLA: In your petition did you
`actually mention those aspects of figure 4? I don't recall
`seeing that in there.
`MR. MESSINGER: In our petition we focused on
`fig. 6 because it was a summary. And this summary reads
`better on that broad, general language that just says, in the
`pixel selection, for example, it says receiving region data and
`then generating one set of pixel data for each region forming a
`
`
`
`13
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`
`
`
`Case IPR2016-00212
`Patent 7,974,339 B2
`
`
`
`new set of data for transmission. This actually shows that
`broad claim language better. So we showed that.
`But it is clear from the overview section of our
`petition in Spriggs, and then the overview of the combination,
`that we are relying on that same algorithm that is in fig. 4.
`And we have additional testimony as well in our reply
`declaration from Dr. Grindon.
`JUDGE HUDALLA: Maybe you could address the
`idea of the pixel data and the region data and whether those
`two aspects are different in Spriggs or is there some common
`data amongst those two things?
`MR. MESSINGER: Yes, Your Honor. In fact, if
`we look at slide 18, with respect to the pixel data, the way we
`refer to it in the petition is we refer to the pixel values, the
`pixels, the pixel data reading on the claim language, and we
`cite that at petition 30, petition page 30, page 31, 44, and 51,
`and then the original filed declaration by Dr. Grindon, 117 to
`118. It is pretty clear that when we refer to pixel data, we are
`referring to the pixel data picture elements in Spriggs.
`For the region data there is language in the petition
`that refers to the corner coordinates for each of the blocks, and
`that corresponds to the region data. And we also identified
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`
`
`
`14
`
`
`
`Case IPR2016-00212
`Patent 7,974,339 B2
`
`
`
`division codes. And I would just, in particular, we mention it
`throughout, but two pages of the petition that you should really
`highlight are petition page 27 and petition page 30.
`We can look at them if you like but those actually
`sort of reflect that we are using the term region data as the
`coordinates of the blocks as well as the division codes that
`identify it, depending on the part of the process.
`JUDGE HUDALLA: Could I ask you a question
`about what we are looking at right now, I guess, under number
`2 there, the third tick point, where it says "blocks as defined
`by corner coordinates and values."
`Is it your contention that the values themselves are
`not part of the region data?
`MR. MESSINGER: Well, as construed by the
`Board, region data is just the coordinates. In the original
`petition we do refer to the coordinates and values in a more
`general way just because that's the entire block and the block
`is identified by the coordinates, and we mention values, but
`when you look at the entire petition, what we are relying on
`under the Board's construction of region data is the
`coordinates.
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`
`
`
`15
`
`
`
`Case IPR2016-00212
`Patent 7,974,339 B2
`
`
`
`But when you look at the original analysis
`selection system, it basically says receiving frame data and
`generating region data, and that's the coordinates of a block.
`JUDGE HUDALLA: So you are saying, has your
`position then changed based on our construction in the
`Decision on Institution?
`MR. MESSINGER: Well, I'm saying what is most
`essential is the corner coordinates. The region data, in terms
`of the values, they are actually present in Spriggs and they are
`present in the algorithm in the way it works, because what
`happens is the corner addresses get stored and there is even a
`description we looked at on slide 11 in Spriggs where it says
`one layer of a stack will have the coordinates, one layer -- and
`then one layer will also have the values.
`We were emphasizing it for completeness just to
`show that the region is really identified by the coordinates.
`The values are present but you do not need to rely on the
`values for region data, and the Board was correct in that.
`In fact, you can see bullet point 3, we sort of
`mention in our reply there is even more discussion of how that
`claimed pixel data is not by itself the region data.
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`
`
`
`16
`
`
`
`Case IPR2016-00212
`Patent 7,974,339 B2
`
`
`
`JUDGE ARBES: Counsel, can you talk a bit about
`the division codes? You also say that the division codes can
`be region data.
`Should we consider the region data to be both then ,
`the division codes and the coordinates, or one or the other?
`What is your position?
`MR. MESSINGER: Yeah, our position is that it
`has to be the coordinates. And in the implementation of
`Spriggs that we are relying on for obviousness, the division
`codes as well. And it is partly because region data is used in
`two elements. So it is used in the analysis to generate the
`region frame, the region data of high and low detail, and that
`can be the corner coordinates.
`And then it also in that step it does generate the
`division coordinates, the division code 1 or zero. And then in
`the pixel selection step it takes that region data, which Patent
`Owner says the pixel selection system is not present, how does
`it do that, but that's completely contrary to the reading of
`Spriggs which says that those coordinates are stored on the
`layer of the stack and then, when there is a subdivision, a
`division code 1, which is properly the -- is transmitted, and
`that reflects the subdivision.
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`
`
`
`17
`
`
`
`Case IPR2016-00212
`Patent 7,974,339 B2
`
`
`
`That's why we call it part of the region data
`because it is yet another identifier that the receiver uses to
`identify the region. And then it also sends the additional
`corner samples. And that is in the pixel selection system.
`And the way the pixels are selected are based on
`the corner addresses that are identified in that layer of the
`stack. The way the receiver reassembles it is using those
`division codes plus corner -- plus coordinates to reassemble the
`block.
`
`JUDGE ARBES: How is the pixel being selected
`based on the region data? Claim 7, for instance, it says it has
`to be based on it. How is the pixel selected based on that?
`MR. MESSINGER: How is it selected based on?
`JUDGE ARBES: On the region data. So you say
`that the coordinates and the division codes are the region data.
`How is the pixel selected based on that?
`MR. MESSINGER: Well, maybe we could look at
`slide 17. The way it is selected based on that is that once there
`is a division of the frame that has been identified to be
`subdivided, then the coordinate addresses of that are stored on
`a stack, and then the division code 1 is transmitted, and then
`
`
`
`18
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`
`
`
`Case IPR2016-00212
`Patent 7,974,339 B2
`
`
`
`the actual corner samples for those new subdivided regions are
`transmitted.
`And they are actually looked up, because it has to
`-- to transmit those corner samples it has to use the coordinate
`addresses of those regions. That's how it knows which corner
`samples to send. And maybe you can look back at fig. 6, if
`you want.
`
`It is laid out in fig. 4 as well, but as our petition
`said for pixel selection system, right under the claim language
`pixel selection we put fig. 6. What gets sent in the pixel
`selection system are those corner samples. So how does that
`rely on the region data like you asked? The only way it can
`get those corner samples SE, SF, SG, is looking at those
`corner -- those coordinate addresses that were stored on that
`stack layer, and then you can look at fig. 4 real quick.
`Fig. 4 says it even more clearly in terms of the
`minutia. It is clear that in fig. 6 it is relying on the addresses
`because that's how it finds the samples. It can't find SE, SF
`unless it looks up the address of SE, SF. And fig. 4 actually
`explicitly says that.
`Fig. 4 of Spriggs, and actually scroll down. So if
`you look at fig. 4, this is what I've been describing. It is
`
`
`
`19
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`
`
`
`Case IPR2016-00212
`Patent 7,974,339 B2
`
`
`
`summarized in fig. 6, fig. 4 has even more details on it, where
`after the interpolation which is above and within the margin,
`there's the -- within the margin is the interpolation step. And
`then once it decides that it has to divide it generates corner
`addresses of the subareas.
`They are pushed onto a stack. And then when it
`transmits the additional corner samples, the way it can do that
`is it is reading those addresses on the stack to sort of send that
`out. I mean, that's kind of how this works, otherwise it
`wouldn't know how to find the corner samples.
`Does that answer your question, Judge Arbes?
`JUDGE ARBES: Yes.
`MR. MESSINGER: And so to answer your
`question as far as construction, our view is that the division
`code is a version of the coordinates in that the system at the
`receiver knows the original coordinates, and then with the right
`division code information it can calculate coordinates. So that
`is what is a little tricky about this.
`If there are no more questions on Spriggs, I will
`move to Belfor in view of Thyagarajan.
`As I mentioned earlier, the essential dispute is
`whether a person of ordinary skill in the art would have
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`
`
`
`20
`
`
`
`Case IPR2016-00212
`Patent 7,974,339 B2
`
`
`
`combined the non-uniform size blocks of Thyagarajan with the
`uniform size blocks of Belfor. And maybe if we look at slide
`20, I will just touch on how Belfor works.
`Belfor discusses advantages, disadvantages of
`using large block sizes, small block sizes, even discusses an
`ideal case where you sort of segment the regions into
`non-uniform regions. But then Belfor does decide to use a
`single block size. And it calls it an important system
`parameter in the example implementation.
`But for what Belfor does it takes that single block
`size and then it uses different modes and sampling analysis to
`vary the frequency of pixels. So that way if you have an image
`that you want, it has a lot of fine detail, and you want to
`sample more heavily, you can use mode 1 and get every pixel.
`If you have an image that is like, say, blue sky with not as
`much spatial variance, then you could use mode 3 and actively
`represent it.
`In some ways Belfor is even smarter than the '339
`patent in that it also continues and has sort of an extended
`discussion about how you can even allocate those modes, in
`that when you are looking at the different blocks and you want
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`
`
`
`21
`
`
`
`Case IPR2016-00212
`Patent 7,974,339 B2
`
`
`
`to say do I want to give it a mode 1 or a mode 2 or mode 3, it
`actually has a process that it goes through.
`And it is a pretty simple process. But what it does
`is it basically uses mode 3. It says, well, let's pick the least
`amount of pixels, which would have the lowest quality image,
`and, if we are still under our bit rate, well, then we can go look
`at other blocks and use mode 2 and actually get a little higher
`quality image as long as we are staying under our bit rate.
`JUDGE HUDALLA: Where is that discussion you
`are referencing right now?
`MR. MESSINGER: It is in the mode allocation
`problem that the Patent Owner references. It is in Belfor. I
`think it's on page -- I would have to look at the page to get you
`the cite. The Patent Owner has an extensive discussion of it.
`I only highlight it here because that mode
`allocation problem is kind of one way in which Belfor
`describes how to allocate modes to meet a desired bit rate. It
`does not -- it is not the kind of allocation problem that would
`prevent anyone from combining Thyagarajan's non- uniform
`size blocks.
`In fact, as we pointed out, if you have non- uniform
`size blocks, you can still allocate modes to those non-uniform
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`
`
`
`22
`
`
`
`Case IPR2016-00212
`Patent 7,974,339 B2
`
`
`
`size blocks. It doesn't preclude someone from combining or
`relying on the non- uniform size blocks of Thyagarajan.
`JUDGE HUDALLA: I have located the mode
`allocation problem section. Does that have any discussion
`about how it is done with non-uniform blocks?
`MR. MESSINGER: No, because Belfor is
`concerned with that example implementation of a common
`single block size when it is talking about the mode allocation
`problem. You are right, Your Honor.
`And if we look at slide 23, just to sort of show how
`the combination of Thyagarajan works so that it would handle
`non-uniform size blocks, what we are relying on is that block
`size assignment 108 highlighted in yellow. That's where the
`subdivision occurs. And fig. 3A teaches the divided image
`block.
`
`And there is actually not that much -- there is not a
`dispute from the Patent Owner, in fact, that Thyagarajan
`teaches divided image block. The dispute comes in terms of
`what information is needed from Thyagarajan to make that
`combination.
`Our position, as laid out in the petition, is that a
`person of ordinary skill in the art, looking at the uniform size
`
`
`
`23
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`
`
`
`Case IPR2016-00212
`Patent 7,974,339 B2
`
`
`
`blocks, and then choosing to now subdivide to get that
`additional level of detail and balance between big and small
`blocks, which is something Belfor mentioned was a concern,
`would know how to use the PQR data. It would know
`Thyagarajan doesn't teach dividing blocks. It also teaches the
`information you need to divide the block.
`And that, for example, is fig. 3C with the PQR
`data. And so with a person combining it, they would just know
`that in Belfor, you send -- if you are sending over uniform
`blocks, well, then you have one block size. But if you are
`using non- uniform size blocks, well, then you need to send
`different block sizes or information on how the blocks are
`divided.
`
`In some ways the PQR data is kind of like the
`division codes we've talked about. It is the bits that show how
`that block is divided. And a person of ordinary skill in the art,
`it wouldn't preclude them from combining it. They would
`know that there is some additional information that is needed
`because now you have non- uniform size blocks, but it's not
`enough that would prevent the combination.
`And Vedanti's own expert even admitted that the
`amount of bits that you would need to sort of identify
`
`
`
`24
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`
`
`
`Case IPR2016-00212
`Patent 7,974,339 B2
`
`
`
`non-uniform size blocks is on the order of six bits or less, you
`know, it is not a lot of information.
`So if we look at slide 28, there is the different
`modes. And even though the claims are such a high level, they
`don't require any kind of particular language that would sort of
`be precluded by a mode allocation, a person of ordinary skill in
`the art, even looking at that mode allocation problem, could
`apply it to the non- uniform size blocks. They would just --
`you could go through and pick the mode 3, the lowest sampling
`frequency, and then, if you had enough bit rate, you could up it
`to mode 2.
`Patent Owner says that this comparison would
`involve a lot of work, but this is within the skill level of a
`person of ordinary skill in the art.
`JUDGE HUDALLA: Do you need the discussion of
`your expert in the supplemental declaration to support that
`contention that it is ordinary skill in the art, because I notice
`that you have quite a lengthy discussion?
`MR. MESSINGER: Yes, we do in terms of the
`particular block size information for non- uniform size blocks,
`if you go to that level of detail with the mode allocation
`problem, but all of that was raised by the Patent Owner.
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`
`
`
`25
`
`
`
`Case IPR2016-00212
`Patent 7,974,339 B2
`
`
`
`What our petition says is what reads on the claims
`themselves. The claims just say generating, receiving region
`data and sort of generating pixel data for transmission. And
`our petition appropriately says that you could combine
`Thyagarajan, which teaches how to subdivide a block, with all
`of the information about how you subdivide the block and have
`bits for that. We rely on that to teach the general -- the
`language needed for the claims.
`When you get down to, well, they couldn't be
`combined because of this particular mode allocation problem,
`that's when we provide additional evidence saying, well, a
`person of ordinary skill in the art wouldn't think that's the kind
`of problem that would prevent a combination. And we
`identified even more information on how it would work for the
`mode allocation with non-uniform size blocks.
`JUDGE ARBES: Counsel, doesn't this argument
`really go to how a person of ordinary skill in the art would
`have meshed the teachings of these two references together?
`Why is that not something that should be in the petition?
`MR. MESSINGER: Your Honor, the petition is
`required to teach why the claims are obvio
Accessing this document will incur an additional charge of $.
After purchase, you can access this document again without charge.
Accept $ Charge
Still Working On It
This document is taking longer than usual to download. This can happen if we need to contact the court directly to obtain the document and their servers are running slowly.
Give it another minute or two to complete, and then try the refresh button.
A few More Minutes ... Still Working
It can take up to 5 minutes for us to download a document if the court servers are running slowly.
Thank you for your continued patience.
This document could not be displayed.
We could not find this document within its docket. Please go back to the docket page and check the link. If that does not work, go back to the docket and refresh it to pull the newest information.
Your account does not support viewing this document.
You need a Paid Account to view this document. Click here to change your account type.
Your account does not support viewing this document.
Set your membership
status to view this document.
With a Docket Alarm membership, you'll
get a whole lot more, including:
- Up-to-date information for this case.
- Email alerts whenever there is an update.
- Full text search for other cases.
- Get email alerts whenever a new case matches your search.
One Moment Please
The filing “” is large (MB) and is being downloaded.
Please refresh this page in a few minutes to see if the filing has been downloaded. The filing will also be emailed to you when the download completes.
Your document is on its way!
If you do not receive the document in five minutes, contact support at support@docketalarm.com.
Sealed Document
We are unable to display this document, it may be under a court ordered seal.
If you have proper credentials to access the file, you may proceed directly to the court's system using your government issued username and password.
Access Government Site
