`
`
`
`
`UNITED STATES PATENT AND TRADEMARK OFFICE
`____________
`
`BEFORE THE PATENT TRIAL AND APPEAL BOARD
`____________
`
`APPLE INC.,
`Petitioner,
`
`v.
`
`QUALCOMM INCORPORATED,
`Patent Owner.
`____________
`
`Cases IPR2018-01250 and IPR2018-01251
`Patent 8,447,132 B1
`____________
`
`Record of Oral Hearing
`Held: October 10, 2019
`____________
`
`
`
`Before TREVOR M. JEFFERSON, DANIEL J. GALLIGAN, and
`AARON W. MOORE, Administrative Patent Judges.
`
`
`
`
`
`
`
`
`
`
`
`

`

`IPR2018-01250 and IPR2018-01251
`Patent 8,447,132 B1
`
`
`
`APPEARANCES:
`
`ON BEHALF OF THE PETITIONER:
`
`
`KIM LEUNG, ESQ.
`TIMOTHY W. RIFFE, ESQ.
`W. KARL RENNER, ESQ.
`Fish & Richardson P.C.
`1000 Maine Avenue, S.W.
`Washington, D.C. 20024
`202-626-6447
`
`
`
`ON BEHALF OF THE PATENT OWNER:
`
`
`EAGLE ROBINSON, ESQ.
`DARREN SMITH, ESQ.
`ERIK JANITENS, ESQ.
`Norton Rose Fulbright US, LLP
`98 San Jacinto Boulevard
`Suite 1100
`Austin, Texas 78701
`
`
`
`
`The above-entitled matter came on for hearing on Thursday, October
`
`10, 2019, commencing at 12:30 p.m., at the U.S. Patent and Trademark
`Office, 600 Dulany Street, Alexandria, Virginia.
`
`
`
`
`
`
`
`
`
`
`2
`
`

`

`IPR2018-01250 and IPR2018-01251
`Patent 8,447,132 B1
`
`
`P R O C E E D I N G S
`- - - - -
`
`USHER: All rise.
`
`JUDGE JEFFERSON: Oh, you may be seated while I get the Judge.
`
`JUDGE GALLIGAN: Good afternoon, this is Judge Galligan. Can
`
`you hear me?
`
`JUDGE JEFFERSON: We can hear you. We -- about a second away
`from seeing you.
`
`JUDGE GALLIGAN: Thank you.
`
`JUDGE JEFFERSON: There we go.
`
`JUDGE GALLIGAN: Good afternoon. I'm Administrative Patent
`Judge Galligan joining from the Texas Regional Office, and before you are
`Judges Jefferson and Moore, and this is a hearing for two IPRs, IPR2018-
`1250 and
`2018-1251 involving U.S. Patent 8,447,132. Petitioner is Apple and Patent
`Owner is Qualcomm. May I have appearances for each side, please? And
`please step up to the podium and make sure the light is green.
`
`MR. RENNER: Okay, yes, sir. Yes, Your Honor, this is Karl Renner
`from Fish & Richardson. I'm joined by colleagues, Tim Riffe and Kim
`Leung, and I guess I'll say it as in before, we'll reserve 30 minutes in terms
`of our direct for redirect. Thank you.
`
`JUDGE JEFFERSON: So 30 minutes for both?
`
`MR. RENNER: Yes.
`
`JUDGE JEFFERSON: Thank you. Patent Owner?
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`
`3
`
`

`

`IPR2018-01250 and IPR2018-01251
`Patent 8,447,132 B1
`
`MR. ROBINSON: Good afternoon, Your Honor, Eagle Robinson for
`
`Patent Owner. With me are Darren Smith and Erik Janitens, and we'd like to
`reserve 20 minutes for surrebuttal, please.
`
`JUDGE GALLIGAN: Thank you. We issued an order in both of
`these cases. We are having one hearing for both cases and each side will
`have 1 1/2 hours of argument, total, so that's 3 hours total of argument time
`for this hearing. Petitioner, you bear the burden of persuasion in showing
`that the challenged claims are unpatentable. You will proceed first; Patent
`Owner may respond. Petitioner, you may have rebuttal time, you reserved
`30 minutes, and Patent Owner, you may have surrebuttal time. With that,
`Petitioner, you may begin.
`
`MR. RENNER: And, Your Honors, locally, can we approach with
`demonstratives?
`
`JUDGE JEFFERSON: Yes. Yes. Thank you.
`
`JUDGE GALLIGAN: Oh, and because I'm remote, please, when
`you're presenting, let me know what slide number you're on, and any other
`paper, please reference explicitly. Thank you.
`
`MS. LEUNG: Yes, Your Honor. May it please the Board, my name
`is Kim Leung and I, along with my colleagues, Karl Renner and Tim Riffe,
`are on behalf of Petitioner Apple, Inc. Two IPRs were instituted against the
`132 Patent, IPR2018-1250 which we'll refer to the 1250 IPR, and IPR2018-
`1251 which we'll refer to as the 1251 IPR. Slide 2, please. So rather than
`walking step-wise through each ground and claim, we'll try to focus in our
`limited time together on a subset of the issues that might benefit from a
`discussion today. For purposes of this discussion, we'll focus on issues 1 to
`4. If the Board would like us to address any particular issue first, or any of
`
`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
`26
`
`4
`
`

`

`IPR2018-01250 and IPR2018-01251
`Patent 8,447,132 B1
`
`the other issues, we can certainly do that; please let me know. Otherwise
`we'll proceed in the order listed here in the Table of Contents.
`
`Slide 5, please. So let's talk a little bit about the 132 patent. So we
`see how the first line of this excerpt from the 132 patent, that the 132 patent
`recognized that techniques for detecting faces and other arbitrary objects and
`patterns and image are known in the art, and you'll also see as we've
`highlighted in this particular slide that the 132 patent acknowledged that
`techniques of dynamic range correction were known, and according to the
`132 patent, though, these techniques of dynamic range correction do not take
`into consideration or use of the content of the image, but the record
`demonstrates that dynamic range correction which considers and uses the
`content of the image was also well known at the time of the 132 patent.
`
`Slide 7. Specifically the 1250 petition which is based on the
`Needham reference establishes that dynamic range correction considers and
`uses the content of the image was well known. Needham is about dynamic
`range correction using the content of the image which are detected image
`features.
`
`Slide 8, please. The 1251 petition which includes grounds based on
`Zhang and Konoplev shows the broad reach of the claims to another type of
`correction, specifically a correction applying different amounts of blurring to
`different portions of an image.
`
`Slide 9. Now that we've provided a brief overview of the 132 patent
`and the grounds and the references, let's go ahead and address the issues. So
`the first issue we have here by patent owner is whether the prior art discloses
`the determined correction is matched to the predetermined type of object
`recited in Claim 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
`26
`
`5
`
`

`

`IPR2018-01250 and IPR2018-01251
`Patent 8,447,132 B1
`
`Slide 10. Now, on the left side is the slide. We've highlighted
`
`portions of Claim 1 that Patent Owner disputes. Specifically Patent Owner
`argues that the prior art doesn't show that the correction is matched because
`the prior art is lacking one, a cause/effect relationship, or, two, a type of
`correction that is predetermined or assigned to a particular image feature.
`Now, even if Patent Owen is correct that the claim to match requires this
`interpretation, Patent Owner is incorrect that the prior art lacks these
`disclosures. So on the next few slides, we'll walk through how the prior art
`discloses a claimed match correction by disclosing either a cause/effect
`relationship or that the correction is predetermined or assigned to the object.
`
`Slide 11. So with respect to the Needham reference, Patent Owner
`acknowledges that Needham discloses matching, but doesn't agree that every
`paragraph of Needham that Petitioner cited in the briefings describes
`matching. So I'm going to go and take you through --
`
`JUDGE GALLIGAN: Counsel, can you tell me where Patent Owner
`said that it agrees that Needham discloses matching?
`
`MS. LEUNG: Yes, so at the 1250 Patent Owner surreply page 3, and
`the 1250 Patent Owner response pages 11, Patent Owner specifically says,
`"When Needham matches a correction to an object in an image," so
`acknowledging that Needham does disclose matching in some portion.
`Needham does not also apply a different amount of the same correction to a
`second portion of the image.
`
`JUDGE GALLIGAN: Thank you.
`
`MS. LEUNG: You're welcome. So I'm going to take you through a
`few portions of Needham that disclose matching. So first here, we have
`portions of Needham that disclose a cause/effect relationship. So as I've
`
`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
`26
`
`6
`
`

`

`IPR2018-01250 and IPR2018-01251
`Patent 8,447,132 B1
`
`highlighted in the top right of the slide here in an excerpt of Needham,
`Needham describes a feature description includes a feature type. Now, the
`feature type indicates a type of object, and Needham gives several examples
`such as a human face, a person, a car, a building, the sky, those are just some
`of the examples that Needham's given. And then in the next excerpt,
`Paragraph 24, Needham describes a correction parameter which defines a
`correction operation and operational parameters used during the correction
`operation. Now, Needham gives examples of correction operations in
`Paragraph 31, which is the next excerpt on the right-hand side, and some of
`those examples of correction operations include increasing the brightness in
`an image, and enhancing the visual contrast in an image, and those are types
`of corrections.
`
`So based on these paragraphs, 30, Paragraph 24, and Paragraph 31 of
`Needham, Needham describes a type of object and a type of correction.
`Needham further says in Paragraph 26 which is the bottom excerpt on the
`right, that based on the feature description the correct unit performs one or
`more specified correction operations. So, in other words, a correction
`operation which is the type of correction is performed based on the feature
`description which includes the feature type. So Needham shows this
`cause/effect relationship between the feature type and the correction
`operation. The feature type, when it's detected, causes a specified correction
`operation to be performed. So there is a cause and effect relationship.
`
`Slide 12. In this slide, we'll show that Needham also meets Patent
`Owner's interpretation that match requires a particular type of correction be
`predetermined or assigned to a particular image feature. Now, Needham
`describes in several paragraphs a correction operation being predetermined
`
`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
`26
`
`7
`
`

`

`IPR2018-01250 and IPR2018-01251
`Patent 8,447,132 B1
`
`for a feature type, so in Paragraph 19, which we've mistakenly cited as
`Paragraph 20, but the top paragraph on the right there, that's Paragraph 19,
`describes a correction operation such as maximizing dynamic range being
`defined on an image feature such as a detected human face. So there's a
`correction type that's defined on a feature type.
`
`Paragraph 20, the next excerpt on the right-hand side, discloses a
`correction operation is defined for each image feature, of multiple image
`features in a single image, so this paragraph is describing when there's one
`image and there is multiple different image features in there, and each image
`feature can have a correction operation defined for it. So, for example, there
`is maximizing dynamic range here described in this paragraph, and that is
`defined for a human face, and then there is also increasing the brightness
`which is defined for the sky.
`
`The next excerpt from Needham, Paragraph 35, indicates that there is
`a corresponding correction operation for each detected image feature, and
`the correction operation is stored in a way that is linked or assigned to the
`correction operation so that it can -- so the correction can be retrieved for the
`image feature.
`
`JUDGE MOORE: So is that corresponding the matching? Where
`exactly is the matching?
`
`MS. LEUNG: Correct. So the corresponding is -- the corresponding
`and the retrieved show that it's matched, because it has to be stored some
`way that there is some link between the image feature and the correction
`operation that the unit knows to retrieve that particular corresponding
`operation for the image feature.
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
`8
`
`

`

`IPR2018-01250 and IPR2018-01251
`Patent 8,447,132 B1
`
`JUDGE MOORE: Well, that looks to me like it's matching features
`
`with correction parameters.
`
`MS. LEUNG: Well, when the feature is matched to the correction
`parameter, then the correction parameter would be matched to the feature.
`
`JUDGE MOORE: So the correction parameter --
`
`MS. LEUNG: The correction parameter includes a -- correction
`parameters includes an operation type, so within the correction parameter
`there -- the correction parameter defines an operation type and also
`operational parameters so when it retrieves a correction parameter, it also
`retrieves a corresponding correction type.
`
`JUDGE MOORE: So the correction types are separate and then the
`matching is between the feature type and the correction parameter which is
`an amount of the correction that is applied?
`
`MR. LEUNG: No, the correction parameter is not separate from the
`correction operation. The correction parameter is -- defines the correction
`operation and also the operational parameters, so it's kind of an umbrella and
`underneath is the type of correction and also the operational parameters
`which indicates the amount of correction, so the correction parameters
`encompasses both the correction type and amount.
`
`JUDGE MOORE: But the feature description and the feature type
`don't, themselves, include the correction?
`
`MS. LEUNG: They don't, themselves, include a correction.
`
`JUDGE MOORE: Okay.
`
`MS. LEUNG: So from these disclosures in Needham, Needham
`shows that the determined correction is predetermined or assigned to a
`feature type.
`
`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
`26
`
`9
`
`

`

`IPR2018-01250 and IPR2018-01251
`Patent 8,447,132 B1
`
`Slide 13. Now, with respect to Nonoka and Konoplev, Patent Owner
`
`disputes that they disclose a claimed matching, but the record has
`established that Nonoka and Konoplev disclose a cause/effect relationship
`between the correction and a type of physical object. So on the left of the
`slide is Figure 8 of Nonoka which is one Nonoka's processing techniques,
`and on the right is a portion of Figure 1 of Konoplev which shows a flow
`diagram of Konoplev's processing technique, or a portion of it, and so I've
`shown in each of Nonoka and Konoplev applies a correction in response to
`detecting the face. So in other words, a correction is triggered based on the
`face being detected. Now, this is shown in Nonoka's Figure 8 where
`Nonoka determines whether a face is detected in step S11, and if the answer
`there is yes, a face is detected, that triggers the optimization processing of
`step S17.
`
`So there, there is a cause/effect relationship where the detection of the
`face causes the optimization processing step S17 to be performed. Now,
`similarly, in Konoplev's Figure 1, when Konoplev finds a face at Step 150, it
`takes the yes branch, and it goes to Step 160 where it processes the facial
`areas to reduce wrinkles. So in other words, a detected face in Konoplev
`triggers wrinkle reduction to be performed, so there's a cause/effect
`relationship there where when the face is detected, it causes the wrinkle
`reduction to be performed.
`
`So in both Nonoka and Konoplev, it doesn't matter that there is a
`correction that was performed before the face was detected, what matters is
`that when the face is detected there is a trigger -- there is a correction that it
`triggers and that correction is applied to the face. Now, this can be the same
`correction that was done previously. It doesn't matter. As long as the face is
`
`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
`26
`
`10
`
`

`

`IPR2018-01250 and IPR2018-01251
`Patent 8,447,132 B1
`
`detected and it triggers a correction whether it's the same correction as
`before or a different correction, then there is a cause/effect relationship
`there. The face causes a correction to be applied. And Claim 1 doesn't
`preclude that there could be some processing that happens before the
`correction is applied to the face.
`
`Slide 14. So the next question here is whether the references disclose
`applying different amounts of the match correction to different portions of
`the image.
`
`Slide 15. So we've highlighted the language here in Claim 1 that
`Patent Owner disputes, and there is no dispute here as to what the
`highlighted claim language means. There is a first amount of the correction
`that is applied to the first portion of the image, and there is a second amount
`of the correction that is applied to the second portion of the image, and the
`amounts applied to the first portion and the second portion are different.
`Nobody disputes that. Patent Owner also doesn't dispute that Nonoka and
`the Zhang/Konaplev combination discloses applying different amounts of
`the correction to different portions of the image.
`
`So with respect to this particular limitation, Patent Owner only
`challenges Needham's disclosure and the combination of Needham and
`Nonoka, so we'll address Patent Owner's challenge to Needham's disclosure
`next, and then the combination of Needham with Nonoka will be addressed
`later when we discuss the motivations to combine.
`
`JUDGE GALLIGAN: And, Counsel, because we're talking about
`both of the cases at the same time, it looks like that's how you're presenting,
`I did have a question on the 1251 case. In our decision on institution on
`pages 11 to 13, we discussed what we considered the two different
`
`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
`26
`
`11
`
`

`

`IPR2018-01250 and IPR2018-01251
`Patent 8,447,132 B1
`
`contentions for Patent Owner -- sorry -- for Petitioner as to how the
`determining steps were done in Claim 1, and the first -- we talked about the
`first way the Petitioner contended was the Zhang feathering process teaches
`determining and applying a correction for the first portion, and we went
`through that and we determined -- we had questions on the sufficiency of
`that showing, and I was looking at the reply. Is Petitioner just focusing that
`on the second of the two on ways that contended which was Konoplev's
`disclosure in combination with Zhang?
`
`MS. LEUNG: Correct. So we are now focusing on that disclosure.
`
`JUDGE GALLIGAN: So for purposes of writing a final, we can
`disregard the first one?
`
`MS. LEUNG: Correct.
`
`JUDGE GALLIGAN: Thank you.
`
`MS. LEUNG: Slide 16. So with respect to Needham's disclosure,
`Patent Owner argues that Needham describes distinct approaches. So what
`Patent Owner argues is that when Needham matches a correction to an
`object, that's implicitly implying that they agree that Needham matches a
`correction to an object. Needham does not also apply a different amount of
`that same correction to a second portion of the image. And so we're going to
`show how Needham does when it matches a correction to an object also
`apply a different amount of that match correction to the first portion and the
`second portion of the image.
`
`So I'm going to take you step-by-step through Figure 8 here of the
`flow chart of Needham, and we're going to be going through this several
`times, so bear with me. So at the top of the corner is Paragraph 35 which
`describes a portion of the flow chart as shown in Figure 8 here, and Figure 8
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`26
`
`12
`
`

`

`IPR2018-01250 and IPR2018-01251
`Patent 8,447,132 B1
`
`shows the steps performed by the correction unit to correct detected features
`in an image. So Paragraph 35 actually relates to the branch where you
`would take to go down the flow chart for correcting individual image
`features. And as we discussed earlier, Paragraph 35 discloses a correction
`being matched to the detected image feature because there
`is -- when an image feature is detected, it retrieves the corresponding
`correction parameter which includes the correction type.
`
`So let's start with going down this flow chart for just generally what
`Paragraph 35 says. So when Needham determines that there is a detected
`feature to correct at Step 840. It's where it says, "More feature to correct."
`It'll go down and retrieve the correction parameter at Step 860 and apply that
`correction type that's in the correction parameter to the image feature, and
`then Steps 860 and 870 are repeated again for each detected image feature in
`the image. So in Paragraph 22 on the right, Needham describes an example
`of how this method of correcting features in a single image would be applied
`so that the same correction is applied to different image features but in
`different amounts.
`
`So in Paragraph 22, there are two types of detected image features.
`There's a human face and there's a building. So let's walk through Figure 8
`here with these two types of image features. So at Step 810, the correction
`unit obtains the feature description and the weight for the detected human
`face, and it obtains a feature description and the weight for the building. So
`as Needham describes in Paragraph 22, the human face feature has a higher
`weight than the building. So at Step 840, the correction unit determines that
`there is a detected face to correct and it proceeds to Step 860 where it
`retrieves the corresponding matched correction parameter, and that
`
`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
`26
`
`13
`
`

`

`IPR2018-01250 and IPR2018-01251
`Patent 8,447,132 B1
`
`correction parameter includes the correction type which in this case is
`dynamic range correction, or maximizing dynamic range, and the correction
`parameter also includes the operational parameter which is the intensity
`dynamic range.
`
`So the correction type that's retrieved here for the face is maximizing
`the dynamic range, and that the parameter that it's using is the intensity
`dynamic range, so how -- what spanned the dynamic range it's going to be
`using.
`
`JUDGE MOORE: So, does feature description include a correction
`parameter for all available corrections?
`
`MS. LEUNG: The feature description for the face?
`
`JUDGE MOORE: I'm sorry?
`
`MS. LEUNG: The feature description for the human face?
`
`JUDGE MOORE: In the end, the feature description for a given
`feature type; does that include a parameter -- sorry -- a correction parameter
`for each available correction type?
`
`MS. LEUNG: The feature description doesn't include any correction
`parameters. The correction parameters are separate from the feature
`descriptions.
`
`JUDGE MOORE: Where are the correction parameters?
`
`MS. LEUNG: The correction parameters are retrieved later, after it
`obtains a feature descriptions.
`
`JUDGE MOORE: And how are they connected to the feature type?
`Maybe I missed something.
`
`MS. LEUNG: The feature descriptions include the feature type.
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
`14
`
`

`

`IPR2018-01250 and IPR2018-01251
`Patent 8,447,132 B1
`
`JUDGE MOORE: Correct. And so where -- how does that connect it
`
`to the parameters?
`
`MS. LEUNG: Based on the feature descriptions, it retrieves the
`corresponding parameters needed to correct that feature type.
`
`JUDGE MOORE: What in the feature description connects it to the
`parameters?
`
`MS. LEUNG: Needham doesn't disclose that there is anything in the
`feature descriptions that connect it to the parameters, but implicitly they
`would have to be linked or assigned somehow so that Needham would know
`that it needs to retrieve a certain operational parameter, correction parameter
`for that feature type.
`
`JUDGE MOORE: Okay.
`
`MS. LEUNG: Okay, so at Step 810 -- I'm sorry -- I forgot where I
`was, so I'll just start here again. At Step 810, the correction unit obtains a
`feature description and a weight for the face which the feature description
`indicates that the detected type is a face. And a feature description and a
`weight for the building and that feature description indicates that the
`detected type is a building. And in Paragraph 22, Needham indicates that
`the human face has a higher weight than the building. So at Step 840, the
`correction unit determines that there is a detected human face and image that
`it needs to correct, so it'll retrieve the correction parameters for it at Step
`860.
`Now, the retrieved correction parameters includes the correction type
`
`which is maximizing dynamic range and it includes the operational
`parameter which is the dynamic range that it's going to be correcting for.
`And at Step 870, it will correct the human face using the retrieved correction
`
`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
`26
`
`15
`
`

`

`IPR2018-01250 and IPR2018-01251
`Patent 8,447,132 B1
`
`type and the operational parameter, and it returns to Step 840 for the
`building feature, and for the building feature it'll similarly performs steps
`860 where it retrieves the corresponding correction parameters which is,
`again, for this particular example, maximizing dynamic range the same one
`as it did for the human face over a dynamic range parameter for the building.
`But in this case, since the -- well, in this case, it'll do the correction
`differently in different amounts because the weight for the face is higher
`than the weight for the building.
`
`So in this case, once Needham has corrected both the face and the
`building features in the image, the face would have gotten a larger correction
`than the building feature because the face had a higher weight than the
`building feature. And so Needham applies different amounts of the same
`correction which was matched to the face to both the face and the building
`feature.
`
`Slide 17. So Patent Owner doesn't agree that Paragraph 22 discloses
`different amounts of correction. Patent Owner says that Paragraph 22 is
`ambiguous and but we disagree. And Dr. Bovik here explains quite clearly
`how Needham applies different amounts of correction to the detected face
`and the detected building. Now, let's take an example here so that we can
`visualize this a little bit more. So in the example of Paragraph 22, Needham
`doesn't say that in this particular example that it's a backlit scene where the
`face is darker than the building or that a flash is used so the face would be
`brighter than the building, so let's assume that the image is evenly lit, and the
`face and the building have the same amount of light on them, the same
`brightness, the same darkness, so let's say they have the same initial dynamic
`range.
`
`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
`26
`
`16
`
`

`

`IPR2018-01250 and IPR2018-01251
`Patent 8,447,132 B1
`
`So taking Dr. Bovik's example in this scenario, the larger dynamic
`
`range used for the face would correspond to a higher increase in the brightest
`value. So let's say the face and the building both start off with the same
`dynamic range, and so since the face has a higher rate, it would get a larger
`increase in the brightest value, so let's say it goes out to here. The building
`having a smaller weight than the face, we're going back to the initial because
`they both have the same initial dynamic range, would have a smaller
`increase in the largest value and the brightest value. And so in this scenario,
`the face would get a larger correction than the builder, and there would be
`different amounts of correction applied to the face and the building.
`
`Slide 18, please. So despite this explanation provided by Dr. Bovik,
`Patent Owner argues that Needham doesn't give the starting dynamic ranges
`of the human face and the building, so we don't know that they get different
`amounts of correction because we don't know what the starting dynamic
`ranges are. Well, Paragraph 22 of course doesn't give the starting dynamic
`range, but I provided one example where if the starting dynamic ranges were
`the same, there would be different amounts of correction, but Needham does
`say that it does take the starting dynamic range into consideration. So in
`specifically Paragraph 32 here, Needham says that the dynamic range may
`be defined as a function of statistical properties of one or more image
`features detected from the input image.
`
`And then in Needham Paragraph 26, which Petitioner has cited in
`other portions of their briefings, Needham states that the statistical properties
`can include maximum and minimum intensity values, which is the starting
`dynamic range. So all this evidence shows that Needham discloses
`determining a correction that is matched to the predetermined type of
`
`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
`26
`
`17
`
`

`

`IPR2018-01250 and IPR2018-01251
`Patent 8,447,132 B1
`
`physical object and applying that match to correction to different portions of
`the image in different amounts. So Needham anticipates Claim 1, and the
`Needham/Nonoka combination renders obvious Claim 1.
`
`Slide 19, please. My colleague, Mr. Riffe, will address issue 3 which
`is whether Konoplev is prior art.
`
`MR. RIFFE: Good afternoon, Your Honors. Again, my name is Tim
`Riffe and I'm on behalf of Petitioner, Apple. As my colleague, Ms. Leung,
`just discussed earlier, Konoplev discloses that the determined correction is
`matched to the predetermined type of physical object. We've also
`established in the record that the Zhang/Konoplev combination teaches
`applying different amounts of the determined correction to different portions
`of the image which Patent Owner did not dispute. However in another
`attempt to defeat the combination of Zhang and Konoplev, Patent Owner
`takes issue with whether Konoplev itself is prior art to the 132 patent.
`
`Well, Konoplev's filing date is after the filing date of the 132 patent;
`we acknowledge that, but the claim's benefit from priority to the earlier filed
`provisional application, and we'll walk through the reasons why. Petitioner
`in our petition in the claim chart in accordance with Dynamic Drinkware, we
`showed where in the provisional application support was provided for
`Konoplev's claims, and we'll touch base on what the challenges are with
`respect to those claims in a few minutes.
`
`Can we go to Slide 21, please. Now, Patent Owner takes issue
`specifically with one specific element of Claims 1 and 13, and that's Element
`E, which recites processing the converted image using wrinkle-reducing
`parameters. Now, first Patent Owner alleges that Konoplev does not provide
`a written description of parameters necessary for wrinkle reduction. Well,
`
`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
`26
`
`18
`
`

`

`IPR2018-01250 and IPR2018-01251
`Patent 8,447,132 B1
`
`let's review briefly what's required to satisfy the written description
`requirement. As Your Honors know, to satisfy the requirement the patent
`specification must describe the claimed invention in sufficient detail that one
`of skill in the art will reasonably conclude that the inventor had possession
`of the claimed invention.
`
`There really should be no question here that the inventor of
`Konoplev's provisional had possession of processing the converted image
`using wrinkle- reducing parameters because as we can see, Konoplev's
`provisional describes specifically processing the image through a smoothing
`filter using such
`wrinkle-reducing parameters. So what does Konoplev's provisional describe
`as the wrinkle-reducing parameters, if we go to Slide 21? The parameters
`include, first, an initial radius, second, a final radius, and they describe
`what's referred to as an imperfection reduction mode which we've
`highlighted on Slide 21, which defines the values for the initial radius and
`final radius. Now, for wrinkle reduction, the parameters would be the
`wrinkle reduction mode, and, again, the corresponding wrinkle-reducing
`radii which they referred to as the initial and final radius.
`
`Slide 22, please. Next Patent Owner argues that Konoplev's
`provisional does not provide an enabling disclosure of using these wrinkle-
`reducing parameters. Now, we've just discussed that Konoplev's provisional
`describes a smoothing filter using filtering parameters including an initial
`radius and a final radius for wrinkle reduction, and here in these paragraphs
`with Konoplev's provision