Ex Parte Brunk et al

Board of Patent Appeals and Interferences

Apr 15, 2009

10045654 (B.P.A.I. Apr. 15, 2009)

UNITED STATES PATENT AND TRADEMARK OFFICE
____________
BEFORE THE BOARD OF PATENT APPEALS
AND INTERFERENCES
____________
Ex parte HUGH L. BRUNK,
BRETT ALAN BRADLEY, and
BRETT T. HANNIGAN
____________
Appeal 2008-2679
Application 10/045,6541
Technology Center 2400
____________
Decided:2 April 15, 2009
____________
Before LEE E. BARRETT, HOWARD B. BLANKENSHIP, and
CAROLYN D. THOMAS, Administrative Patent Judges.

BARRETT, Administrative Patent Judge.

DECISION ON APPEAL

1 Filed October 26, 2001, titled "Including a Metric in a Digital
Watermark for Media Authentication."
2 The two-month time period for filing an appeal or commencing a
civil action, as recited in 37 C.F.R. § 1.304, begins to run from the decided
date shown on this page of the decision. The time period does not run from
the Mail Date (paper delivery) or Notification Date (electronic delivery).
Appeal 2008-2679
Application 10/045,654

2
This is a decision on appeal under 35 U.S.C. § 134(a) from the final
rejection of claims 15-19, 23 and 25-27. Claims 1-9 and 11-14 are allowed.
Claims 10 and 24 are canceled. Claims 20-22 are objected to for their
dependency on a rejected claim. We have jurisdiction pursuant to
35 U.S.C. § 6(b).
We reverse.
STATEMENT OF THE CASE
The invention
The invention relates to methods of steganography and data hiding,
which can be used to help authenticate media. In particular, the methods
include a watermark having various properties, such as being designed to be
lost or degrade upon signal processing, and having a metric used for media
authentication.

The claims
Claim 25 is illustrative:
25. A digital watermarking method comprising:
embedding a digital watermark in a media signal, the digital
watermark being designed to be lost or to degrade upon at least one
form of signal processing;
determining a metric for the embedded digital watermark, the
metric comprising a benchmark for the embedded digital watermark;
embedding the metric in the media signal; and
Appeal 2008-2679
Application 10/045,654

3
embedding data in the media signal, the data indicating how the
metric was determined.

The reference

Shur US 6,330,672 B1 Dec. 11, 2001

The rejections
Claims 15-19, 23, and 25-27 stand rejected under 35 U.S.C. § 102(e)
as being anticipated by Shur.3

DISCUSSION
Claims 25-27
Appellants present three reasons why the Examiner erred. We are
persuaded by each of the reasons.

3 Although the statement of the rejection in the (Third) Examiner's
Answer (3d Ans.) entered January 16, 2008, includes claim 22 in the
anticipation rejection, and includes an obviousness rejection of
claims 20 and 21 over Shur and Takaragi, it is elsewhere stated that
"claims 20-22 are objected to for their dependency on a rejected base claim"
(3d Ans. 8), which is consistent with the first and second Examiner's
Answers. It appears that these rejections were erroneously included in
response to a appeal return notice on January 2, 2008, and did not intend to
reinstate the rejection. Thus, claims 20-22 are objected to.
Appeal 2008-2679
Application 10/045,654

4
1.
Issue
Does Shur teach "embedding a digital watermark in a media signal,
the digital watermark being designed to be lost or to degrade upon at least
one form of signal processing," as recited in independent claim 25?

Contentions
The Examiner finds that Shur teaches a "watermark having the
capacity to degrade to prevent use of the digital data after a predetermined
date or number of plays (column 10, lines 1-11)" (Final Rej. 8; 3d Ans. 6).
Appellants discuss that examples of watermarks designed to be lost or
to degrade are a "semi-fragile" digital watermark which degrades in
response to some types of degradation of the watermarked signals but not
others, and a "fragile" digital watermark where the strength (or other signal
characteristics) of the watermark signal in a copy of the watermarked object
is less than the strength in the original object (Br. 9). It is argued that, in
Shur, "[t]here is no discussion of how the permitted number of plays will be
used to degrade the watermark or cause the watermark to become lost"
(id. at 10).
The Examiner states that changing the permissible number of plays in
the watermark "degrades" the watermark with each play (3d Ans. 8). It is
stated that claim 25 does not recite the terms "fragile" or "semi-fragile" (id.).

Appeal 2008-2679
Application 10/045,654

5
Facts
The portion of Shur cited by the Examiner describes that "parametric
data may comprise . . . [a] number of permitted plays or other parameter
relating to the expiration of any license or lease" (col. 10, ll. 1-9).

Analysis
We agree with Appellants that embedding a watermark containing
data about a permitted number of plays describes nothing about the
watermark being lost or degraded upon signal processing, as claimed. Even
if the watermark is changed with each play to account for the number of
plays, we disagree with the Examiner's interpretation that a change in the
number of plays "degrades" the watermark. The watermark is not lost or
degraded, the information in it is just changed to a different number. The
watermark is what carries the information, not the information itself.

Conclusion
Shur does not teach "embedding a digital watermark in a media
signal, the digital watermark being designed to be lost or to degrade upon at
least one form of signal processing." For this reason, the anticipation
rejection of independent claim 25 and dependent claims 26 and 27 is
reversed.

Appeal 2008-2679
Application 10/045,654

6
2.
Issue
Does Shur teach "determining a metric for the embedded digital
watermark, the metric comprising a benchmark for the embedded digital
watermark," as recited in independent claim 25?

Contentions
The Examiner finds that Shur teaches "[g]enerating a metric based on
the detected watermark (column 11, lines 12-24)" (Final Rej. 8; Ans. 6).
Appellants argue that the claim limitation is "determining a metric,"
not "generating a metric" as stated by the Examiner and, in any case,
denying use of recorded data based on the watermark fails to teach
determining a metric, the metric comprising a benchmark (Br. 10).
The Examiner states that the metric is the number of plays, which is
determined from data within the watermark, and the benchmark is the
number of permissible plays, in this case zero (3d Ans. 9).

Facts
The Specification describes that an authentication metric is used to
help measure or gauge the degradation in the watermark (para. [0083]), such
as watermark signal strength, a watermark correlation value, a ratio of
values, etc. (para. [0084]).
Appeal 2008-2679
Application 10/045,654

7
The cited portion of Shur teaches extracting the watermark and
denying use of the recorded data based on data in the watermark. Extracting
the watermark would require access to a key, the random number sequence
and the perceptual coding function used to encode the watermark.

Analysis
We do not agree with the Examiner's interpretation that a number of
plays is a "metric for the embedded digital watermark" and the number of
plays remaining being zero is a "benchmark for the embedded digital
watermark." While the number of plays is a "metric" in the broadest sense
of a "standard of measure" of something, the number of plays has nothing to
do with measuring the watermark: it is not a metric "for" the watermark.
The fact that the system will deny use of the recorded data when a particular
datum in the watermark reaches the number zero again has nothing to do
with the watermark itself: the datum is not a benchmark "for" the watermark.

Conclusion
Shur does not teach "determining a metric for the embedded digital
watermark, the metric comprising a benchmark for the embedded digital
watermark." For this second reason, the anticipation rejection of
independent claim 25 and dependent claims 26 and 27 is reversed.

Appeal 2008-2679
Application 10/045,654

8
3.
Issue
Does Shur teach "embedding data in the media signal, the data
indicating how the metric was determined," as recited in independent
claim 25?

Contentions
The Examiner finds that Shur teaches "[e]mbedding said metric into
[the] data stream as part of the digital watermark (column 10, lines 1-11 and
column 8, line 56-column 9, line 4)" (Final Rej. 8; 3d Ans. 6).
Appellants argue that the rejection does not address the limitation in
question (Br. 10-11).
The Examiner finds that Shur teaches "[e]mbedding of the digital
watermark into the media stream includes indicia of the metrics that
constitute the digital watermark (Shur, column 4, lines 13-34)" (3d Ans. 9)
where "indicia refer to the type of transaction that is purchased by the user,
namely a least of media stream, and the limitations of that transaction, the
number of permitted plays" (id.).
Appellants respond that the Examiner maps the term "metric" to the
number of permitted plays, but Shur does not indicate how the permitted
number of plays was determined (Reply Br. 4). It is also argued that the
information related to a lease and a permitted number of plays does not
include data indicating how a metric is determined.
Appeal 2008-2679
Application 10/045,654

9
Facts
The Specification describes embedding data indicating the method
used to determine the metric, which is used to recover the metric: "a digital
watermark includes a payload (or message) indicating how metric 20 was
calculated. In this alternative, metric module 32 is versatile, comprising the
ability to generate metric 34 according to the specified metric 20 generation
criteria." Para. [0091].
Shur describes that the embedded watermark may contain "indicia of
the form of the transaction" (col. 4, ll. 25-26), such as the expiration date of
a lease or "data indicating the number of licensed plays" (col. 4, ll. 29-30).
Shur describes that embedded "watermark input parametric data may
comprise . . . [a] number of permitted plays or other parameter relating to the
expiration of any license or lease" (col. 10, ll. 1-9).

Analysis
"Indicia of the form of the transaction," of the types described in Shur,
are not a "metric comprising a benchmark for the embedded digital
watermark" because the indicia deals with information unrelated to the
watermark. We also agree with Appellants that "indicia of the form of the
transaction," assuming it is a "metric," says nothing about "how the metric
was determined," as recited in claim 25. It is not sufficient that data is
embedded in the media signal, the data must be a metric and must specify
how it was determined.
Appeal 2008-2679
Application 10/045,654

10
Conclusion
Shur does not teach "embedding data in the media signal, the data
indicating how the metric was determined." For this third reason, the
anticipation rejection of independent claim 25 and dependent claims 26 and
27 is reversed.

Claims 19 and 23
Appellants present two reasons why the Examiner erred.
1.
Issue
Does Shur teach "embedding second information in the media signal,
the second information corresponding to a rendering channel through which
the media signal will be rendered," as recited in independent claim 19?

Contentions
The Examiner finds that Shur describes "[e]mbedding the encoding
algorithm for the digital data in the media stream (column 10, lines 1-11)"
(Final Rej. 8; 3d Ans. 5).
Appellants argue that "while the cited encoding algorithm may relate
generally to an act of embedding, it is not understood to teach or suggest a
rendering channel, e.g., a printer, a streaming device or broadcast
channel" (Br. 13).
Appeal 2008-2679
Application 10/045,654

11
The Examiner states that it is implicit "that the media stream is meant
to be played (rendered) in some kind of signal reproduction device"
(3d Ans. 10) and when the media is put in the device, the "device 'reads' the
digital watermark and if the watermark is valid plays the media signal" (id.).

Facts
The Specification describes:
Metric 20 is adjusted to accommodate the degradation prediction.
Degradations may be a function of the media signal (12 or 16) or of an
anticipated rendering process. For example, a prediction or estimate
of how the watermark 10 degrades, e.g., as a function of media signal
12's local characteristics and/or using knowledge of the anticipated
legitimate media rendering system (e.g., a printer, broadcaster, etc.) is
included in metric 20.

Para. [0086].
The portion of Shur cited by the Examiner states that "watermark
input parametric data may comprise . . . the encoding algorithm that was
used to encode the signal via perceptual coder 110" (Shur, col. 10, ll. 1-10).

Analysis
We agree with Appellants that the encoding algorithm for encoding
the signal has nothing to do with the rendering channel, i.e., with how the
media will be reproduced or represented. The fact that media has to be
rendered on some device as stated by the Examiner does not imply the
Appeal 2008-2679
Application 10/045,654

12
existence of information corresponding to the rendering channel through
which the media signal will be rendered. For example, digital media on a
compact disc does not need to include information identify the rendering
channel because the user knows to put it in a CD player.

Conclusion
Shur does not teach "embedding second information in the media
signal, the second information corresponding to a rendering channel through
which the media signal will be rendered." For this reason, the anticipation
rejection of independent claim 19 and dependent claim 23 is reversed.

2.
Issue
Does Shur teach "analyzing the digital watermark embedded in the
media signal to determine a baseline state for the digital watermark," as
recited in independent claim 19?

Contentions
The Examiner finds that the "baseline state" corresponds to the
number of permitted plays (Final Rej. 7; 3d Ans. 5).
Appellants argue that "carrying this position forward, if a number of
permissible plays is determined from watermarked media . . . , information
corresponding to this number must then be embedded in the watermarked
Appeal 2008-2679
Application 10/045,654

13
media signal" (Br. 13), but there is no suggestion of providing this extracted
number back to an embedder for re-embedding (id. at 14).
The Examiner refers to data that limits the permissible number of
plays for the rendering channel. The Examiner responds that there is no
limitation of re-embedding the extracted information. The data is checked to
see if the digital watermark would result in a digital signals not protected by
a watermark (3d Ans. 11).
Appellants argue that the Examiner's new position impermissibly
relies on the number of permitted plays for both "the first information
corresponding to the baseline state of the digital watermark" and "the second
information corresponding to a rendering channel through which the media
signal will be rendered" (Replay Br. 5-6).

Facts
We do not find the term "baseline" in the Specification. Appellants
refer to paragraphs [0087] and [0088] and Figures 4a and 4b for this
limitation (Br. 5). We interpret "analyzing the digital watermark embedded
in the media signal to determine a baseline state for the digital watermark" to
correspond to "a metric analysis of the unaltered watermarked media
signal 22" (para. [0088]) where the "baseline state" is the "unaltered" signal.

Appeal 2008-2679
Application 10/045,654

14
Analysis
Under the Examiner's interpretation, the number of permitted plays is,
perhaps, a "baseline" for the number of times a media can be played, but we
do not see how it can be fairly described as a "baseline state for the digital
watermark." Again, the information in the watermark is different than
information about the watermark itself. Thus, the Examiner's interpretation
is not persuasive.
Appellants' arguments about re-embedding are construed to mean that
the watermark must somehow keep track of how many times the media has
been played to determine whether the number of plays has exceeded the
baseline. It is not clear how this argument relates to the claim limitation.
However, Appellants' argument that the Examiner has shifted position to
rely on the number of plays for both the "first information" and the "second
information" is persuasive because the same data cannot satisfy both
limitations. For this additional reason, the Examiner's reading of the claim
onto Shur is not persuasive.

Conclusion
Shur does not teach "analyzing the digital watermark embedded in the
media signal to determine a baseline state for the digital watermark." For
this additional reason, the anticipation rejection of independent claim 19 and
dependent claim 23 is reversed.

Appeal 2008-2679
Application 10/045,654

15
Claims 15-18
Appellants present two reasons why the Examiner erred.

1.
Issue
Does Shur teach "embedding a digital watermark in a media signal,
the digital watermark being designed to be lost or to predictably degrade
upon predetermined signal processing," as recited in independent claim 15?

Conclusion
We find that Shur does not teach this limitation for the reasons stated
in the discussion of claim 25. For this reason, the anticipation rejection of
independent claim 15 and dependent claims 16-18 is reversed.

2.
Issue
Does Shur teach "generating a metric based on the detected digital
watermark," as recited in independent claim 15?

Contentions
The Examiner finds that Shur teaches "[g]enerating a metric based on
the detected watermark (column 11, lines 12-24)" (Final Rej. 7; 3d Ans. 4).
Appeal 2008-2679
Application 10/045,654

16
The cited portion of Shur teaches extracting the watermark and
denying use of the recorded data based on the watermark.
Appellants argue that they fail to see how this teaches generating a
metric as recited in claim 15 (Br. 16).
The Examiner states that the metric is the permitted number of plays,
which has to be generated to be understandable to the media reproduction
device (3d Ans. 12).

Facts
The portion of Shur cited by the Examiner teaches extracting the
watermark and denying use of the recorded data based data in the
watermark, such as the number of plays.

Analysis
It appears that Shur merely reads the permitted number of plays from
the watermark. It is not clear that the permitted number of plays is a
"metric" within the ordinary meaning of a "standard of measure" because it
is simply a number recorded in the watermark. Assuming the permitted
number of plays is a "metric," it is not clear that it is "generated" in the
ordinary meaning of "to bring into existence." See In re Scroggie, 170 Fed.
Appx. 132, 135 (Fed. Cir. 2006) (nonprecedential) ("The term 'generating
page data' means that the page data is 'generated,' not merely 'selected.'").
Appeal 2008-2679
Application 10/045,654

17
We do not agree with the Examiner's finding that reading a number
embedded in the watermark is "generating a metric."

Conclusion
Shur does not teach "generating a metric based on the detected digital
watermark." For this additional reason, the anticipation rejection of
independent claim 15 and dependent claims 16-18 is reversed.

CONCLUSION
The rejection of claims 15-19, 23, and 25-27 is reversed.

REVERSED

rwk

DIGIMARC CORPORATION
9405 SW GEMINI DRIVE
BEAVERTON, OR 97008