Ex Parte Yu et al

Patent Trial and Appeal Board

Nov 14, 2016

12170476 (P.T.A.B. Nov. 14, 2016)

UNITED STA TES p A TENT AND TRADEMARK OFFICE
APPLICATION NO. FILING DATE
12/170,476 07/10/2008
74365 7590 11/16/2016
Slater Matsil, LLP
17950 Preston Road, Suite 1000
Dallas, TX 75252
FIRST NAMED INVENTOR
Lu Yu
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
ATTORNEY DOCKET NO. CONFIRMATION NO.
HW 0710114US 1148
EXAMINER
HESS, MICHAEL J
ART UNIT PAPER NUMBER
2481
NOTIFICATION DATE DELIVERY MODE
11/16/2016 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):
uspatent@huawei.com
docketing@slatermatsil.com
PTOL-90A (Rev. 04/07)
UNITED STATES PATENT AND TRADEMARK OFFICE
BEFORE THE PATENT TRIAL AND APPEAL BOARD
Ex parte LU YU, CIXUN ZHANG, QICHAO SUN,
and LIANHUAN XIONG
Appeal2016-000106
Application 12/170,476
Technology Center 2400
Before JEAN R. HOMERE, KARA L. SZPONDOWSKI, and
AARON W. MOORE, Administrative Patent Judges.
MOORE, Administrative Patent Judge.
DECISION ON APPEAL
Appeal2016-000106
Application 12/170,476
STATEMENT OF THE CASE
Appellants 1 appeal under 35 U.S.C. § 134(a) from a Final Rejection of
claims 22-34, which are all of the pending claims. We have jurisdiction
under 35 U.S.C. § 6(b ).
We affirm.
THE INVENTION
The application is directed to an "interpolating method and a device
for interpolating in grade video compression." (Abstract.) Claim 22,
reproduced below, is representative:
22. A method for performing interpolation in scalable video cod-
ing, which is used for interlayer prediction in scalable video cod-
ing, the inter-layer prediction comprising performing interpola-
tion filtering processes on luminance components and chromi-
nance components by using independently different interpolation
filters with the same pel accuracy,
wherein; in use of an Extended Spatial Scalability (ESS) tech-
nology, the interpolation filters comprise:
4-tap interpolation filters with a 1/16-pel accuracy for the lu-
minance components, and 2-tap interpolation filters with a 1/16-
pel accuracy for the chrominance components;
wherein the 2-tap interpolation filters with a 1/16-pel accu-
racy are in a form of {Fo1, Fo2}/[Fo1 + Fo2], Poi being the tap
coefficients of the filters, where i = 1 ... 2, where Poi is the adjust-
ment of Fot' according to Poi= Poi' x 2j or round (Foi' x '.Y) ± 1,
where i = 1 ... 2, and j is an arbitrary integer.
1 Appellants identify Huawei Technologies Co., Ltd. as the real party in
interest. (See App. Br. 2.)
2
Appeal2016-000106
Application 12/170,476
THE REFERENCES
The prior art relied upon by the Examiner in rejecting the claims on
appeal is:
Cismas
Sato
US 2004/0190632 Al
US 2004/0213470 Al
Sept. 30, 2004
Oct. 28, 2004
E. Francois et al., Generic Extended Spatial Scalability
(Proposal JVT-0041 ), Joint Video Team of ISO/IEC MPEG &
ITU-T VCEG, 13th Meeting: Palma de Mallorca, Spain,
October 18-22, 2004 ("Francois")
Sheng Zhong et al., Improved Sub-Pel Interpolation Filters
(Proposal VCEG-N 51 ), ITU - Telecommunications
Standardization Sector, Video Coding Experts Group (VCEG),
14th Meeting: Santa Barbara, CA, USA, September 24--27,
2001 ("Zhong")
THE REJECTIONS
1. Claims 22-25, 27-29, and 31-33 stand rejected under 35
U.S.C. § 103(a) as unpatentable over Francois, Cismas, Zhong, and Sato.
(See Final Act. 7-10.)
2. Claims 26, 30, and 34 stand rejected under 35 U.S.C. § 103(a)
as unpatentable over Francois, Cismas, Zhong, Sato, "and further in view of
tools and knowledge available to those skilled in the art." (See Final Act.
11-12.)
APPELLANTS' CONTENTIONS
Appellants argue that the rejections are in error for the following
reasons:
1. The "combination fails to teach or suggest '4-tap interpolation
filters with a 1/16-pel accuracy for the luminance components, and 2-tap
3
Appeal2016-000106
Application 12/170,476
interpolation filters with a 1/16-pel accuracy for the chrominance
components,' as recited in claim 22." (See App. Br. 5-9.)
2. The Examiner fails to establish a prima facie case of
obviousness with regard to the specific tap coefficients recited in claims 32,
33, and 34. (See App. Br. 10-14.)
ANALYSIS
Claim 22
The Examiner found that although Francois does not teach "4-tap
interpolation filters with a 1/16-pel accuracy for the luminance components,
and 2-tap interpolation filters with a 1/16-pel accuracy for the chrominance
components," such an arrangement would have been obvious in light of the
combination of Francois, Cismas, Zhong, and Sato. (See Final Act. 7-8.)
Specifically, the Examiner found as follows:
As those references teach in combination, many tap/phase com-
binations can be applied to pixel interpolation. Given the well-
known phenomenon that human visual perception is more sensi-
tive to luminance components than chrominance components,
and as Cismas teaches, it is appropriate to apply smaller, simpler
tap/phase configurations to chrominance than to luminance. This
approach is obvious to one skilled in the art striving to achieve
an optimal balance between the best interpolation (best picture)
and limited computational resources. In fact, practitioners in the
field faced with limited resources often devote a greater percent-
age of resources to luminance components.
(Final Act. 8.) The Examiner further explained that "Cismas discloses a 6-
tap filter for the luminance component and a 2-tap filter for the
chrominance components" such that it "teaches independently different
interpolation filters as in claim 22" and that "although Cismas does not
explicitly teach ... using the same pel accuracy ... using the same pel
4
Appeal2016-000106
Application 12/170,476
accuracy is merely an obvious design choice." (Final Act. 3--4.) The
Examiner further explained that it would have been obvious to employ the
well-known 4:4:4 sampling scheme (i.e., no down-sampling), which would
have resulted in the same accuracy for each filter. (Id. at 4.)
Appellants argue that "Cismas fails to disclose that the different filters
have the same pel accuracy and the pel accuracy is 1/16-pel accuracy."
(App. Br. 7.) Appellants further argue that "the Examiner's assertion is
untenable, because no convincing evidence was provided by the Examiner to
support the assertion that the interpolation pel accuracy relies on the
sampling schemes." (App. Br. 7 .)
The Examiner acknowledges that "it may seem as though Examiner
shifted away from Appellants' claims regarding interpolation filter accuracy,
toward an unrelated discussion of chrominance subsampling (sampling
ratios)," but notes that "the two concepts are closely intertwined." (Ans.
10.2) The Examiner further cites Dang (US 2006/0133506 Al; published
June 22, 2006), finding it to suggest that a 4:4:4 sampling ratio would
suggest equal interpolation precision. (Id. at 11.) Appellants do not respond
to that finding.
On this record, we conclude that Appellants have not shown error in
the Examiner's analysis. In particular, we agree that it would have been
obvious to one of ordinary skill in the art at the time of the invention to
employ a sampling ratio that would have resulted in the luminance and
2 Certain comments by the Examiner at this point in the Answer and
elsewhere in the prosecution stray from the appropriate decorum. The
Examiner is reminded to conduct all USPTO business in a professional
manner.
5
Appeal2016-000106
Application 12/170,476
chrominance filters having the same accuracy. Appellants argued the
Examiner's finding that a known 4:4:4 sampling ratio would have resulted in
equal accuracy was not supported by evidence (see App. Br. 7), the
Examiner provided further support in the Answer (see Ans. 11 ), and
Appellants did not reply. We also agree with the Examiner that the degree
of accuracy (e.g., 1/16 pel) would have been an obvious matter of design
choice, not invention. The Specification does not ascribe any particular
significance to that degree of accuracy, and Appellants do not identify any in
the briefing.
For these reasons, we sustain the Section 103 rejections of claims 22-
24, 27-29, and 31.
Claims 25, 26, 30, 32, 33, and 34
Appellants argue that these claims are non-obvious due to their
recitations of specific coefficients. (See App. Br. 10-14.) In particular,
Appellants argue that "the Examiner erred by using the 'obvious to try
rationale,"' because:
The specific tap coefficients are derived from considerable
effort, which requires the inventor to consider many factors that
might influence the quality of the filter, for example, the phase,
the application scenario, the texture of an image frame in base
layer and so on, and to tune the coefficients of each filter
scrupulously, and finally a group of candidates of coefficients
that were obtained. The candidates of coefficients then went
through a series of contesting experiments and finally a best one
was identified. All of those experiments surely follow some
general principle/guide, such as the Lanczos filter design guide,
but guide is not enough to derive such a set of specific tap
coefficients without any inventive action, even when considering
the teaching of the combination of any of the cited references.
(App. Br. 11.)
6
Appeal2016-000106
Application 12/170,476
With respect to claim 26, the Examiner explained that "fairly straight-
forward MATLAB code (or similar mathematical software) can generate
such coefficients for the tap/phase combinations listed in the claim" and one
of ordinary skill in the art would have been motivated to do that "because
such tools are part of the general knowledge available to those skilled in the
art and because it is preferable to leave complex, computationally intensive
mathematical calculations to computer programs." (Final Act. 11.) The
Examiner further cited a reference "disclosing 5-tap, 16 phase filter
coefficients generated using a Matlab script." (Id.)
We are not persuaded by Appellants' discussion of the "considerable
effort" involved in arriving at the claimed coefficients because it is only
attorney argument, which may not take the place of evidence. See, e.g., In
re De Blauwe, 736 F.2d 699, 705 (Fed. Cir. 1984). Instead, we agree that
"the prior art teaches the coefficients are balanced or weighted according to
position" and that, "[ t ]herefore, there are only a small handful of possibilities
for coefficients to try and the skilled artisan would be able to experiment
with this very small number of possibilities to fit the coefficients to desired
outcomes." (Ans. 16.) Again, Appellants do not provide evidence sufficient
to overcome this finding. We are also unpersuaded by Appellants' argument
concerning MATLAB because, as the Examiner explains (see Ans. 16-17),
it was identified for illustrative purposes only and does not form a basis for
the rejection. Appellants do not argue or provide evidence that MATLAB or
similar mathematical software was not well known.
We thus agree that, on this record, the selection of the particular
coefficients claimed would have been an obvious matter of routine
7
Appeal2016-000106
Application 12/170,476
experimentation and, accordingly, sustain the Section 103 rejections of
claims 25, 26, 30, 32, 33, and 34.
DECISION
The rejections of claims 22-34 are affirmed.
No time period for taking any subsequent action in connection with
this appeal may be extended under 3 7 C.F .R. § 1.13 6( a )(1 )(iv).
AFFIRMED
8