IEEE Consumer Electronics Magazine - July 2015 - 85

previous invention had implemented
because all of the previous algorithms
were implemented in a computer after
the image was captured. If you think
about it, this is a pretty important and
essential step in deciding if you should
apply a red-eye filter to an image.
Also note that this claim is somewhat
shorter than your average pencil. (If you
check the prosecution record, this patent
application was originally rejected and
only subsequently granted after an
appeal process; thus, the publication
date is nearly five years after the original
application filing. But, as the pencil test
is only an urban legend, one shouldn't
read too much into this.)

THE SCOPE OF CLAIMS
It is important to understand the human
nature of the IP process. Patent attorneys argue on behalf of their clients,
but often the underlying invention is
modest in nature or the search process
uncovers additional prior art that anticipates key elements of the invention.
On the other hand, a patent examiner's goal is to process as many cases as is
practical; few cases can be rejected out of
hand, and, if the applicant's attorney is
tenacious and the client has deep pockets,
it is often a long road to an outright rejection. The allowance of a narrow claims
scope brings the prosecution to a conclusion but allows the patent attorney to
impress his/her client with a grant of patent. In this way, both the examiner and
attorney often find common ground.
There are different forms that these
claims of limited scope take, but after a
while, you learn to spot them. We've
already discussed those lengthy claim
constructions that have created the
urban legend of the pencil test. Here
are some other examples.

CLAIMS THAT DON'T
MAKE SENSE
One variant takes the form of patents
with allowed claims that simply do not
make sense. As an example consider:
...wherein the video processing
circuit combines the R channel
and the B channel from a normal
video output and obtains a grayscale of or higher than 14 bits.

Those skilled in image analysis
know that the green (G) channel provides the most accurate and sharpest
luminance/gray scale images; for this
very reason, conventional image sensors dedicate 50% of their sensors to
green pixels, so this requirement loses
50% of the useful video data.
The reality is that this is a bad way to
obtain a grayscale image, and any conventional approach will use the G channel. The patent is, naturally, perfectly
valid and enforceable-if you can find
anyone using the protected technique.

CLAIMS WITH MATH FORMuLAS
You'll often see a patent claim that
includes a math formula or a matrix
structure or the equivalent. These are
often converted into text form, appearing
somewhat incomplete in the main U.S.
Patent and Trademark Office text database. The full formula will be preserved,
however, in the original PDF document.
Here is an example:
...conducting XOR on a descriptor
vector mQj q corresponding to a
jth minutia of a query fingerprint

CLAIMS THAT ARE
CONTRAINDICATED
Sometimes you find claims that simply
can't work. Remember that a key requirement for the grant of a patent is nonobviousness, i.e., a technique that goes against
standard teaching in a field of knowledge.
Really good inventions tend to have a
strong element of this-often they are
due to mistakes or accidental events that
demonstrate that something contrary to
conventional wisdom does work.
But you also have to remember that
you can be granted a patent on something that simply does not work. After
all, patents are legal documents, not
engineering ones. A patent examiner is
only responsible for the legal correctness of the patent claims. So, something
that does not work in practice can still
meet the necessary criteria for the grant
of a patent. (Technically speaking, one
of the grounds for rejecting a patent is
that it is nonoperable, and an examiner
can request that a working embodiment
be shown in operation; however, in practice, this is rarely requested.) Perhaps it
is time to return to the Chinese puzzle.

CONCLuSION

The allowance of a narrow
claims scope brings the
prosecution to a conclusion
but allows the patent
attorney to impress his/her
client with a grant
of patent.
image and auxiliary data {ei,
h(ci)} corresponding to an ith
minutia of a fingerprint template,
i.e., ci′ = ei5mQj^q.
So, why is math bad? The reality is
that any enforcement proceedings are
going to be held in a courtroom and
argued by lawyers in front of a judge, and
mathematical expressions are somewhat
abstract. In this case, demonstrating
infringement becomes a much more
complex task. Claims that involve simpler, tangible concepts-e.g., thresholds,
parameters, textures, or conditions-are
considerably easier to demonstrate in a
legal proceeding.

There I was, wondering whether it was
nobler in the mind to suffer the slings
and arrows of outrageous fortune
(remember that quick opinion?) or to
take arms against a sea of (machinetranslated) claim sets and, by analyzing,
end them. (Apologies to Hamlet fans.)
Yes, I had screen after screen of
these patents to review, but fortunately
I had a few slings and arrows of my
own. A quick check with the nearest
pencil eliminated a significant volume
of my troubles; a closer inspection of
several others confirmed my suspicions
that many public research institutions
are sometimes overly proud of the
complexity of their research achievements and their mathematical heritage.
Fortified by a strong cup of coffee,
your hero next scrutinized the remaining handful of more athletic and finetuned troubles, finally arriving at a
suitable quick opinion.

(continued on page 89)
july 2015

IEEE ConsumEr ElECtronICs magazInE

Table of Contents for the Digital Edition of IEEE Consumer Electronics Magazine - July 2015