Ex Parte LI et al

Patent Trial and Appeal Board

Dec 14, 2017

14125157 (P.T.A.B. Dec. 14, 2017)

United States Patent and Trademark Office
UNITED STATES DEPARTMENT OF COMMERCE
United States Patent and Trademark Office
Address: COMMISSIONER FOR PATENTS
P.O.Box 1450
Alexandria, Virginia 22313-1450
www.uspto.gov
APPLICATION NO. FILING DATE FIRST NAMED INVENTOR ATTORNEY DOCKET NO. CONFIRMATION NO.
14/125,157 12/10/2013 Wenlong Li P53394US 1758
88032 7590 12/18/2017
TnrHanTPT aw T T C
EXAMINER
12501 Prosperity Drive, Suite 401
Silver Spring, MD 20904
CHANG, KENT WU
ART UNIT PAPER NUMBER
2624
NOTIFICATION DATE DELIVERY MODE
12/18/2017 ELECTRONIC
Please find below and/or attached an Office communication concerning this application or proceeding.
The time period for reply, if any, is set in the attached communication.
Notice of the Office communication was sent electronically on above-indicated "Notification Date" to the
following e-mail address(es):
info@jordaniplaw.com
admin @jordaniplaw.com
inteldocs_docketing @ cpaglobal. com
PTOL-90A (Rev. 04/07)
UNITED STATES PATENT AND TRADEMARK OFFICE
BEFORE THE PATENT TRIAL AND APPEAL BOARD
Ex parte WENLONG LI, THOMAS SACHSON, and YUNZHEN WANG
Appeal 2017-007372
Application 14/125,157
Technology Center 2600
Before NATHAN A. ENGELS, JAMES W. DEJMEK, and
MICHAEL M. BARRY, Administrative Patent Judges.
BARRY, Administrative Patent Judge.
DECISION ON APPEAL
Appellants1 appeal under 35 U.S.C. § 134(a) from a Final Rejection of
claims 18—42, which are all of the pending claims. Claims 1-17 have been
canceled. App. Br. 6. We have jurisdiction under 35 U.S.C. § 6(b).
We affirm-in-part.
1 Appellants identify Intel Corp. as the real party in interest. App. Br. 3.
Appeal 2017-007372
Application 14/125,157
Introduction
Appellants describe the invention as “relating] to animating an avatar
based on facial motion data, such as user facial motion data corresponding to
one or more facial gestures observed when a user emulates a source object.”
Spec. ^ 1. Claim 18 is exemplary:
18. An apparatus to animate an avatar, comprising:
an image capture device;
a script module to present script data to a user that is to
include text taken from speech made in a performance by a
source object in media content;
a motion module to extract user facial motion data that
corresponds to one or more user facial gestures observed by the
image capture device when the user is to utilize the script data to
verbalize the text taken from the speech made in the performance
by the source object; and
an avatar module to animate an avatar based on the user
facial motion data.
App. Br. 25 (Claims App’x); see also Spec. 10-23, Figs. 1A-B.
Rejections and References
1. Claims 18, 20-24, 26-35, and 37-42 stand rejected under
35 U.S.C. § 103 as unpatentable over Jones et al. (US 2012/0106796 Al;
pub. May 3, 2012, hereinafter “Jones”) and Poggio et al. (US 6,250,928 Bl;
iss. June 26, 2001, hereinafter “Poggio”). Final Act. 2-13.
2. Claims 19, 25, and 36 stand rejected under 35 U.S.C. § 103 as
unpatentable over Jones, Poggio, and Wang et al. (US 2010/0082345 Al;
pub. Apr. 1, 2010, hereinafter “Wang”). Id. at 13-14.
2
Appeal 2017-007372
Application 14/125,157
ISSUES
Based on Appellants’ arguments, the issues are whether the Examiner
errs in:
(a) finding the combination of Jones and Poggio teaches or suggests
“a script module to present script data to a user that is to include
text taken from speech made in a performance by a source object
in media content,” as recited in independent claim 18, and as
recited commensurately in independent claims 23 and 34 (App.
Br. 11-13);
(b) finding the combination of Jones and Poggio teaches or suggests
“a motion module to extract user facial motion data that
corresponds to one or more user facial gestures observed by the
image capture device when the user is to utilize the script data to
verbalize the text taken from the speech made in the performance
by the source object,” as recited in independent claim 18, and as
recited commensurately in independent claims 23 and 34 (id. at
13-14);
(c) combining Jones and Poggio for the rejections of independent
claims 18, 23, and 34 because
(1) the combination would lead away from the claimed
invention,
(2) modifying Jones as necessary based on the teachings of
Poggio would require a substantial reconstruction and redesign
and a change in the basic principles of operation of Jones, or
(3) the Examiner failed to provide an articulated reasoning with
rational underpinning (id. at 18-23);
(d) finding the combination of Jones and Poggio teaches or suggests
“emulat[ing] a corresponding source object when the
corresponding source object in the media content speaks in the
performance,” as recited in claim 22 (id. at 14-16);
(e) finding the combination of Jones and Poggio teaches or suggests
“wherein the user audio is to be generated by the verbalization of
the text by the user,” as recited in claim 26 and commensurately
recited in claim 37 (id. at 16-17);
3
Appeal 2017-007372
Application 14/125,157
(f) finding the combination of Jones and Poggio teaches or suggests
“to extract the source facial motion data from media content that
includes one or more of a music video or an animated movie in
which the user is absent, wherein the source object is to include
one or more of a performer or a character in the media content,” as
recited in claim 28 {id. at 17-18).
ANALYSIS
Claims 18, 23, and 34
Claim Interpretation —Non-Functional Descriptive Material
As an initial matter, we interpret the phrases “taken from speech made
in a performance by a source object in media content” and “taken from the
speech made in the performance by the source object,” as recited in
claim 18, and as commensurately recited in claims 23 and 34. When a claim
element merely describes the nature of information, it may be directed to
nonfunctional descriptive material, which does not lend patentability to an
otherwise unpatentable claim. See In re DiStefano, 808 F.3d 845, 848-51
(Fed. Cir. 2015); Ex parte Nehls, 88 USPQ2d 1883, 1889 (BPAI 2008)
(precedential) (holding “the nature of the information being manipulated
does not lend patentability to an otherwise unpatentable computer-
implemented product or process”); see also Ex parte Curry, 84 USPQ2d
1272, 1274 (BPAI 2005) (informative) (finding the fact that data was
“wellness-related” did not functionally affect other claim requirements and,
thus, constituted nonfunctional descriptive material that could not render
nonobvious an otherwise obvious claim), affd, No. 06-1003 (Fed. Cir.
2006) (Rule 36).
Claim 18 recites that the source of text of the script module’s script
data is “taken from speech made in a performance made by a source object
in media content” without any requirement that functionally relates the
4
Appeal 2017-007372
Application 14/125,157
source of the script data to the other requirements of the script module. The
Specification, which discloses that “script data [] may include performance
information which may direct the actions of the user [] in a karaoke fashion”
(T| 16) and that a user “may, in one example, use the script data [] to emulate
the performance of the source object” (]f 22), does not disclose that the
source of the script data must be functionally related to the script module.2
Claim 18 also recites a motion module that verbalizes text “from the
speech made in the performance by the source object.” Similar to the script
module, the recited requirements for the motion module include no
functional limitations related to the source of the text. The Specification
does not di sclose that the source of the text functionally affects the motion
module’s extraction of user facial motion data.
Accordingly, the phrases “taken from speech made in a performance
made by a source object in media content” and “taken from the speech made
in the performance by the source object,” as recited, are nonfunctional
descriptive material and, thus, these phrases cannot be used to distinguish
claims 18, 23, and 34 over the prior art. Ex parte Nehls, 88 USPQ2d at
1889; cf. In reNgai, 367 F.3d 1336, 1338-39 (Fed. Cir. 2004).
The Combined Teachings of Jones and Poggio
Appellants contend the Examiner errs because Jones and Poggio
individually fail to teach or suggest “a script module,” as recited in claim 18.
App. Br. 11-13; Reply Br, 5-7. In particular, Appellants assert “Jones
2 We also note Appellants state embodiments “relate to animating an avatar
based on facial motion data, such as user facial motion data corresponding to
one or more facial gestures observed when a user emulates a source object.”
Spec. ^ 1 (emphasis added).
5
Appeal 2017-007372
Application 14/125,157
merely teaches to use capture devices 302 to capture detectable features 114
of the user and not a source object in media content, much less that script
data is presented to the user that includes text taken from speech made in a
performance by a source object in media content.'''’ App. Br. 12; see also
id. at 11-13; Reply Br. 5-7. Appellants further assert “Jones does not teach
a source object in media content, such as an actor in a movie that a user
attempts to emulate, as claimed,” but “at most merely teaches a user in a
field of view of a capture system 110 to create a custom avatar.” Reply
Br. 6. Appellants also assert Poggio “merely teaches to use a set of key
words to create a photo-realistic talking face and not that a set of key words
are presented to the human subject that includes text taken from speech
made in a performance by the photo-realistic talking face.” App. Br. 12.
Appellants’ arguments are unpersuasive. To the extent Appellants’
argue Jones does not teach “taken from speech made in a performance made
by a source object in media content,” as recited, we are unpersuaded
because, as discussed above, it is nonfunctional descriptive material.
Appellants do not persuasively rebut the Examiner’s finding that the
combined teachings of Jones and Poggio teach “a script module to present
script data to a user that is to include text. . .,” as required by claim 18. See
Final Act. 2-4 (citing Jones 1| 60; Poggio 2:44-55). The Examiner finds,
and we agree, by disclosing presentation of training scripts to a user to
prompt the user to respond in a particular manner, Jones teaches “a script
module to present script data to a user,” as recited. Final Act. 2-3 (citing
6
Appeal 2017-007372
Application 14/125,157
Jones ^ 60).3 The Examiner additionally finds Poggio teaches or at least
suggests “to include text taken from speech,” because Poggio’s method of
producing photo-realistic talking face images is based on recordings of a
human subject’s facial responses when enunciating one or more specifically
designed sets of key words. Final Act. 3—4 (citing Poggio 2:44-55); see also
Poggio 3:1-6, 6:21-25. The Examiner reasons it would have been obvious
to combine the teachings of Jones and Poggio because the combination
provides “the ability of the final audio-visual output to look and sound
exactly as if it were produced by a real human face.” Final Act. 4.
Appellants do not persuade us of error in the Examiner’s findings and
reasoning.
Appellants further contend the Examiner errs because Jones and
Poggio individually fail to teach or suggest “a motion module,” as recited in
claim 18. App. Br. 13..14: Reply Br. 7-8. In particular, Appellants assert
Jones teaches using capture devices to capture detectable features of the
user, but does not teach extracting facial motion data “when the user is to
utilize script data that includes text taken from speech made in a
performance by a source object in media content, much less to extract when
the user is to utilize script data to verbalize text taken from speech made in a
performance by the source objectApp. Br. 14 (emphasis modified).
Appellants further assert Poggio merely teaches using a set of key words to
create a photo-realistic talking face, but does not teach extracting facial
3 By disclosing training scripts that “prompt a user to speak particular words
or phrases,” Jones alone teaches “a script module to present script data to the
user that is to include text,” as recited. Jones ^ 59.
7
Appeal 2017-007372
Application 14/125,157
motion data when verbalizing “text taken from speech made in a
performance by the photo-realistic talking face.” Id.
These arguments are unpersuasive because, as discussed above,
“taken from the speech made in the performance by the source object,” as
recited, is nonfunctional descriptive material. The Examiner finds, except
for the requirement “to verbalize the text,” Jones teaches the recited “motion
module” by disclosing capturing facial expression and movement of a user
in response to presenting the user a training script to create a custom avatar
that reflects the user’s distinguishable attributes. Final Act. 3 (citing Jones
"fli 32, 55, 60). The Examiner also finds Poggio teaches “to verbalize the
text” by disclosing converting input text into speech for a talking face image.
Id. at 4 (citing Poggio 2:44-55, 65-67). Appellants do not persuade us the
Examiner errs in these findings that Jones and Poggio teach the “motion
module” requirements, as recited.
Appellants additionally argue the Examiner’s proposed combination
of Jones and Poggio is erroneous because (1) the combined teachings would
lead away from the claimed invention, (2) the modification of Jones with
Poggio would require a substantial reconstruction and redesign and a change
in the basic principles of operation of Jones, and (3) the Examiner failed to
provide an articulated reasoning with rational underpinning. App. Br. 18-
23; Reply Br. 9..15.
Appellants’ “lead away” arguments are unpersuasive because they
simply point out that there are differences between Jones, Poggio, and the
claimed invention, which does not constitute a teaching away. See In re
Beattie, 974 F.2d 1309, 1312-13 (Fed. Cir. 1992). A teaching away requires
a reference to actually criticize, discredit, or otherwise discourage the
8
Appeal 2017-007372
Application 14/125,157
claimed solution. See In re Fulton, 391 F.3d 1195, 1201 (Fed. Cir. 2004).
Appellants present no such evidence of teaching away in either Jones or
Poggio.
We also disagree with Appellants that modifying Jones with Poggio
would require a substantial reconstruction, redesign, or a change in Jones’s
basic principles of operation. Jones’s training scripts can prompt the user to
speak a particular word, which teaches the very elements for which the
Examiner relies on Poggio—i.e., “to include text” and “to verbalize the
text.” See, e.g., Jones ]}]} 58, 62.
Finally, contrary to Appellants’ assertion that the Examiner failed to
articulate reasoning with rational underpinning, the rejection specifically
articulates a reasonable rationale to combine the teachings of Jones and
Poggio—i.e., to provide “the ability of the final audio-visual output to look
and sound exactly as if it were produced by a real human face.” Final
Act. 4. In other words, modifying Jones to include the teachings of Poggio
would provide for an avatar that emulates a real human face. Poggio
supports this rationale by producing an animated avatar that accurately lip
syncs corresponding speech while performing natural facial and body
motions. Poggio]} 43. Jones also supports such a need for improvement in
the accuracy of avatars. Jones ]} 5. Appellants do not persuasively rebut the
Examiner’s proffered rationale.
Accordingly, we sustain the rejection of claim 18, and also of claims
19-21, 23, 25, 27, 29, 30, 33, 34, 36, 38^10, and 42, for which Appellants
present no arguments separate from those for claim 18.
9
Appeal 2017-007372
Application 14/125,157
Claims 22, 24, 31, 32, 35, and 41
Claim 22 depends from claim 18 and adds, inter alia, the requirement
“to emulate a corresponding source object when the corresponding source
object in the media content speaks in the performance.” App. Br. 26
(Claims App’x). In rejecting claim 22, the Examiner finds Jones teaches this
limitation by its disclosure of an avatar creator system that receives
differential attributes which it applies to create a custom avatar that reflects
the distinguishable user attributes described in the differential attributes.
Final Act. 6 (citing Jones 27-30). Appellants argue “Jones merely
teaches to receive multiple sets of user attributes [] and not to emulate
anything, much less to emulate a source object in media content or that each
user of a plurality of users is to emulate a corresponding source object.'''’
App. Br. 15. In response, the Examiner explains that Jones teaches or
suggests a musical performer that the user tries to emulate. Ans. 7 (citing
Jones ^ 23).
We agree with Appellants. Unlike claim 18, in which there is a lack
of any functional requirement for the script module or motion module based
on the text in the script data coming from a particular “source object,” claim
22 functionally requires source object emulation “when the corresponding
source object in the media content speaks in the performance.” Paragraphs
27-30 of Jones do not mention emulating a source object speaking in a
performance, but instead merely disclose receiving and applying differential
attributes of a user to a base avatar to create a custom avatar that reflects
distinguishable user attributes. Jones ^ 27-30. Paragraph 23 of Jones does
not mention any type of performance of a source object in media content,
but instead merely discloses that detectable features of a user may include
10
Appeal 2017-007372
Application 14/125,157
structural, external, and behavioral features, including words spoken by a
user or voice patterns of user speech. Id. 23.
Thus, the Examiner has not provided evidence or technical
explanation for how Jones teaches or suggests “to emulate a corresponding
source object when the corresponding source object in the media content
speaks in the performance,” as recited. Further, although the cited portions
of Poggio discloses converting input text into an audio-visual speech stream,
resulting in a talking face image enunciating the text, the Examiner has not
shown that Poggio teaches or suggests the disputed limitation. See Final
Act. 3—4 (citing Poggio 2:44-55, 65-67). Nor has the Examiner provided an
adequate rationale to fill the gaps in the cited prior art. See Final Act. 5-6;
Ans. 5-8.
Accordingly, we do not sustain the Examiner’s § 103 rejection of
claim 22. For the same reasons, we do not sustain the rejections of claims
24, 31, 32, 35, and 41, each of which recites a similar limitation and stands
similarly rejected based on the disclosures of Jones and Poggio.
Claims 26 and 37
Appellants argue the Examiner errs in rejecting claims 26 and 37
because Jones and Poggio fail to teach verbalizing “text taken from speech
made in a performance by a source object.” App. Br. 16-17; see also Reply
Br. 8-9. In other words, Appellants argue claims 26 and 37 based on
requirements recited in their respective parent independent claims 23 and 34.
Accordingly, because, as discussed above, the source of the text as recited in
claims 23 and 34 is non-functional descriptive material, we sustain the
Examiner’s rejection of claims 26 and 37.
11
Appeal 2017-007372
Application 14/125,157
Claim 28
Claim 27, from which claim 28 depends, adds the requirement for
“the computing device to extract source facial motion data that corresponds
to the source object of the script data and to animate the avatar based on the
source facial motion data” to independent claim 23. App. Br. 27 (Claims
App’x). Claim 28 further requires “the computing device to extract the
source facial motion data from media content that includes one or more of a
music video or an animated movie in which the user is absent, wherein the
source object is to include one or more of a performer or a character in the
media content.” Id. (emphasis added). Thus, the recited extraction of facial
motion data in claim 28 is functionally limited by requirements recited for
media content. In other words, unlike independent claims 18 and 23, in
which, as discussed above, the claim recitation related to “media content” is
non-functional descriptive matter, claim 28 functionally requires extracting
facial motion data from the recited media content in which the user is absent.
The Examiner finds Jones’s disclosure of a “[sjtereoscopic image
capture system [that] includes multiple video or other image capture devices
which capture image streams from multiple angles and other types of sensors
to detect outlines of objects” teaches the added requirement of claim 28.
Final Act. 9 (citing Jones Fig. 3, item 320). Appellants argue the Examiner
errs because “Jones and Poggio are completely silent regarding media
content that includes a music video and/or an animated movie, much less in
which a user is absent, or that a source object may include a performer
and/or a character in media content, as claimed.” Reply Br. 9; see also
App. Br. 17-18. Instead, Appellants assert Jones merely teaches to elicit an
12
Appeal 2017-007372
Application 14/125,157
emotional response such as a humored laugh with a picture or a scenario.
App. Br. 17.
We agree with Appellants. The Examiner does not explain how
Jones’s stereoscopic image capture system, which Jones discloses is used for
extracting the user’s facial attributes for creating a customized avatar (see,
e.g., Jones 2, 48-60), teaches claim 28’s requirement to extract facial
motion data from media content in which the user is not present. There is no
finding that Poggio teaches or suggests the disputed limitation, and the
Examiner does not provide an adequate rationale to fill the gaps in the cited
prior art. See Final Act. 6-7, 8-9; Ans. 5-8.
Accordingly, we do not sustain the rejection of claim 28.
DECISION
For the above reasons, we affirm the § 103 rejections of claims 18-21,
23, 25-27, 29, 30, 33, 34, 36^40, and 42, and we reverse the § 103 rejections
of claims 22, 24, 28, 31, 32, 35, and 41.
No time period for taking any subsequent action in connection with
this appeal may be extended under 37 C.F.R. § 1.136(a)(l)(iv).
AFFIRMED-IN-PART
13