Ex parte FUJISAKI et al.

Board of Patent Appeals and Interferences

May 14, 1998

08061495 (B.P.A.I. May. 14, 1998)

Application for patent filed May 12, 19931
1
THIS OPINION WAS NOT WRITTEN FOR PUBLICATION
The opinion in support of the decision being entered today
(1) was not written for publication in a law journal and
(2) is not binding precedent of the Board.
Paper No. 19
UNITED STATES PATENT AND TRADEMARK OFFICE
__________
BEFORE THE BOARD OF PATENT APPEALS
AND INTERFERENCES
__________
Ex parte TETSUNOSUKE FUJISAKI, WILLIAM D. MODLIN,
KOTTAPPURAM M. MOHIUDDIN, and HIROYASU TAKAHASHI
__________
Appeal No. 95-4730
Application 08/061,4951
__________
ON BRIEF
__________
Before KRASS, JERRY SMITH, and BARRETT, Administrative Patent
Judges.
KRASS, Administrative Patent Judge.
DECISION ON APPEAL
Appeal No. 95-4730
Application No. 08/061,495

2
This is a decision on appeal from the final rejection of
claims 1 through 20, constituting all the claims in the
application.
The invention is directed to a system for and method of
handwriting recognition. More particularly, both dynamic and
static (the static being obtained by processing the dynamic)
handwriting information are obtained and these two types of
information are employed to obtain recognition results which
are then merged from both to obtain a most probable recognized
character.
Representative independent claim 1 is reproduced as
follows:
1. A handwriting recognition system, comprising:
handwriting transducer means, responsive to a handwriting
input from a user, for outputting corresponding time ordered
stroke information;
first means, having an input coupled to an output of said
handwriting transducer means, for determining a first list
comprised of at least one probable character that the
corresponding time ordered stroke information is intended to
represent;
means, having an input coupled to said output of said
handwriting transducer means, for converting the corresponding
time ordered stroke information to static stroke information;
second means, having an input coupled to an output of
said converting means, for determining at least one second
Appeal No. 95-4730
Application No. 08/061,495

3
list comprised of at least one probable character that the
static stroke information represents; and
means, having a first input coupled to an output of said
first determining means and a second input coupled to an
output of said second determining means, for merging said
first list and said at least one second list to provide a
third list comprised of at least one most probable character
that the corresponding time ordered stroke information is
intended to represent.
The examiner relies on the following references:
Tyburski et al. Re. 31,692 Oct. 2, 1984
(Tyburski)
Holt 4,837,842 Jun. 6, 1989
Fujisaki 5,029,223 Jul. 2, 1991
Guyon et al. 5,105,468 Apr. 14, 1992

(Guyon)
Claims 1 through 20 stand rejected under 35 U.S.C. § 103.
As evidence of obviousness, the examiner cites Fujisaki, Holt
and Tyburski with regard to claims 1 through 3, 5 through 9,
11 through 14 and 17 through 20, adding Guyon to this
combination with regard to claims 4, 10, 15 and 16.
Reference is made to the briefs and answer for the
respective positions of appellants and the examiner.
OPINION
We reverse.
Appeal No. 95-4730
Application No. 08/061,495

4
The examiner offers Fujisaki to show a handwriting
transducer responsive to a handwriting input for outputting
time ordered stroke information and establishing a first list
of at least one probable character that the time ordered
stroke information is intended to represent. The examiner
then offers Holt as disclosing a character recognition system
which forms
static representations of on-line handwritten input in order
to extract characteristic data. This much is not disputed by
appellants.
However, the examiner recognizes that even though
Fujisaki and Holt are evidence of individually forming,
respectively, time ordered stroke information (dynamic) and
static representations of handwritten input, the instant
claims also require a merging of the two lists provided by
these systems. Therefore, the examiner relies on Tyburski as
a linking reference, ostensibly for the proposition that “it
has long been recognized in the art of character recognition
that the combination of complementary recognition algorithms
can achieve improved accuracy results” [answer-page 4].
Appeal No. 95-4730
Application No. 08/061,495

5
To the extent that the examiner is relying on, and taking
official notice of, common knowledge in the art that “the
combination of complementary recognition algorithms can
achieve improved accuracy results,” we cannot agree absent any
evidence showing what the examiner alleges to be well known.
Assertions of technical facts in areas of esoteric technology,
such as the handwriting recognition technology before us, must
always be supported by citation to some reference work
recognized as standard in the pertinent art and the appellant
given the opportunity to challenge the correctness of the
assertion or the notoriety or repute of the cited reference.
In re Ahlert, 424 F.2d 1088, 1091, 165 USPQ 418, 420 (CCPA
1970).
The only evidence proffered by the examiner as proof of
the alleged common knowledge in the art is the Tyburski
reference. Thus, the question is whether or not Tyburski
does, indeed, teach that “the combination of complementary
recognition algorithms can achieve improved accuracy results.”
After review of the Tyburski reference, we find that it
does disclose the use of two disparate types of character
recognition, one being magnetic, the other being optical. If
Appeal No. 95-4730
Application No. 08/061,495

6
the magnetic reader fails to recognize a character while the
optical reader does recognize the character, than the
character is identified as that corresponding to the character
recognized by the optical reader. Conversely, where the
magnetic reader recognizes a character while the optical
reader fails to do so, the character is identified as that
corresponding to the character recognized by the magnetic
reader. In the situation where both readers provide signals
indicative of different characters, either one or the other is
chosen, depending on parameters of the system, or,
alternatively, a reject signal is generated. Thus, it appears
that Tyburski chooses one or the other or neither, but does
not choose in accordance with some combination or “merging” of
the data, in the sense of appellants’ disclosed merging
wherein the lists generated by the time ordered information
and the static information are physically merged into a third
list which is then used to choose the most likely character.
The examiner does make a cogent argument [answer-page 9]
that since a list, as claimed, need only contain one
character, a final selection of one character based on two
derived characters would meet the claim requirement.
Appeal No. 95-4730
Application No. 08/061,495

7
Accordingly, one might argue that Tyburski is “merging” a
first and second list to provide a third list by choosing
either the result of the magnetic reader or the result of the
optical reader and finally selecting one of those results. Of
course, one might also argue that there is no “merging,” as
disclosed and claimed by appellants, in Tyburski because
Tyburski chooses one, or no, result of two distinct character
input devices and never physically combines these results in
any manner. In any event, we do not reach the question of
whether Tyburski, in fact, teaches “merging,” as broadly
claimed, because we find other claim limitations which are not
disclosed or suggested by the applied references.
First, the two types of character recognition systems,
besides being two distinct systems, detect the same static
characteristics of printed characters. Therefore, since both
the magnetic and optical character recognition devices of
Tyburski function as static character recognition devices,
there is no suggestion in Tyburski of combining lists
comprised of both static stroke information and dynamic, or
time ordered stroke information. Thus, the question arises as
to why the skilled artisan would have employed such a teaching
Appeal No. 95-4730
Application No. 08/061,495

8
in Tyburski to then combine the time ordered system of
Fujisaki with the static information system of Holt in any
manner. We find no suggestion to do so. Tyburski does not
disclose a combination of “complementary recognition
algorithms,” if the examiner means by this that there are
disclosed in Tyburski both time ordered and static information
relative to character recognition. Tyburski discloses only
static information. Therefore, there is no teaching or
suggestion by Tyburski that “complementary recognition
algorithms can achieve improved accuracy results,” as
contended by the examiner.
Further, each of the instant independent claims requires
the static stroke information to be derived from the time
ordered stroke information. As depicted in instant Figure 1a,
for example, the feature vector representing the time ordered
stroke information is input to box 16. Simultaneously
therewith, the time ordered stroke information is input to
replicator 18 and into converter 20 wherein the time ordered
stroke information is converted to static stroke information.
The time ordered stroke information and the static stroke
information are then processed in parallel to arrive at two
Appeal No. 95-4730
Application No. 08/061,495

9
candidate lists which are then merged at 24 to result in a
final recognition result.
There is nothing in the applied references which, in any
way, suggests deriving the static stroke information from the
time ordered stroke information and the examiner has never
addressed this issue in the answer. Accordingly, the examiner
has not established the requisite prima facie case of
obviousness required for a proper rejection under 35 U.S.C. §
103.
The Guyon reference was applied against dependent claims
4, 10, 15 and 16 for a showing of neural networks for
performing character recognition. However, Guyon does not
provide for any of the deficiencies noted supra with regard to
the other three applied references as applied against the
independent claims. Accordingly, we will not sustain the
rejection of any of the claims under 35 U.S.C. § 103.
The decision of the examiner rejecting claims 1 through
20 under 35 U.S.C. § 103 is reversed.
Appeal No. 95-4730
Application No. 08/061,495

10
REVERSED
Errol A. Krass )
Administrative Patent Judge )
)
)
)
Jerry Smith ) BOARD OF
PATENT
Administrative Patent Judge ) APPEALS AND
) INTERFERENCES
)
)
Lee E. Barrett )
Administrative Patent Judge )
Appeal No. 95-4730
Application No. 08/061,495

11
Harry F. Smith
Perman & Green
425 Post Road
Fairfield, CT 06430