Recognicorp, LLC v. Nintendo Co. -- Petition for En Banc Rehearing and Amicus Briefs

By Paul Cole* --

Proceedings for infringement of U.S. Patent No. 8,005,303 (Recognicorp, assigned from IQ Biometrix) resulted in an appeal decided on 28 April 2017, which decision was reviewed in this space by Michael Borella, and also criticised (see "Regnoicorp -- A Miscarriage of Justice Calling for En Banc Reconsideration").

As expected, a petition for en banc rehearing was filed on 30 May, and was followed on 11 June by an Independent Inventors' amicus brief, and on 13 June by a further amicus brief by the author.

Recognicorp explained that encoding and decoding data using computers is an important and innovative field of technology and objected that the panel opinion threatened the validity of all encoding and decoding patents, including patents covering important technologies such as MP3 audio coding, MPEG video coding, JPEG image coding and encryption.  It cited decisions where such claims have been upheld including DDR Holdings, LLC v. Hotels.com; Cal. Inst. of Tech. v. Hughes Communs., Inc.; and TQP Dev and LLC v. Intuit Inc.

Their brief identified at least two inventive concepts recited in the claims of the '303 patent:

(1) using a single composite facial image code composed of a number of facial feature element codes, such as for lips or eyes, to represent the entire facial image instead of encoding the image pixel-by-pixel as was done in the prior art; and

(2) using a particular algorithm to combine the various facial feature element codes together into a single composite facial image code by multiplying the facial code by a code factor for each facial feature.

They argued that these qualified as improvements in technology or in a technical field.  There was no basis for holding that the fields of image encoding and data compression were not "technical field[s]" in which patentable improvements could be made, nor could reducing bandwidth and memory storage requirements be disregarded but instead enjoyed status as specific improvements in the functioning of a computer.  The improvements achieved here should be contrasted with Digitech where no improvement was asserted.

Although the panel opinion objected that, like Digitech, "a user starts with data, codes that data using 'at least one multiplication operation,' and ends with a new form of data" Recognicorp replied that this was plainly not what the claims of the '303 patent recited.  They did not recite "start[ing] with data," or even "gathering . . . data [without] . . . input from a physical device," but rather began with "displaying facial feature images" which are "select[ed]" "via a user interface."  Similarly, the claims did not "end with a new form of data," but rather they ended with the display, on a second display, of the reproduction of a facial image based on the composite facial image code.  There was no basis for finding the claims ineligible under Alice step one merely because of the involvement of a mathematical method, compare MCro v Bandai, and also Enfish, which started with a mathematical table of rows and columns, performed mathematical operations and ended with a self-referential table for a computer database.

In relation to the second step of Alice, it was argued that the panel wrongly rejected the potential that mathematical equations and algorithms, even when implemented on a computer to produce tangible benefits in that computer's operation, could constitute patentable inventive concepts.  The panel opinion failed to perform any such analysis -- it did not attempt to analyze whether the "facial feature element codes," "code factors," and the "particular encoding process using the specific algorithm disclosed" that used those codes to create a "composite facial image code" was "significantly more" than "the abstract idea of encoding and decoding image data".  Recognicorp argued that improvements in computer and software technology often take the form of new and innovative mathematical operations and algorithms, which when appropriately claimed as being performed on a computer as part of a technological process, should be eligible for patent protection.

The Individual Inventors' brief also took the position that the panel opinion had ignored the rule in Alice that eligibility under § 101 requires an analysis of all claim limitations as an ordered combination.  That rule had been clarified in McRO to require that all claim limitations be addressed in step 1 of the Alice test but had been discarded by the decision.  The wisdom "that courts 'must be careful to avoid oversimplifying the claims' by looking at them generally and failing to account for the specific requirements of the claims" stated in McRO had been ignored.  The consequence was that under the instant decision, patent examiners not only now have carte blanche to ignore any claim limitation and issue of evidence under step 1 of Alice, but are free to ignore decades of established precedential decisions focused on preemption in favor a per se rule having no basis in preemption.

The Cole brief followed many of the arguments of Recognicorp and the Individual Inventors and asked:

• What scope is attributable to the phrase "directed to" in the first stage of the Alice test, at what level of abstraction is interpretation of a representative claim impermissibly untethered from the express language of that claim, and is an abstraction permissible where it omits reference to elements positively recited in a representative claim?

• Is a new and beneficial result improving a technological process available to rebut an objection that claimed subject matter falls within the law of nature, natural phenomenon or abstract idea exception?

• Is the Alice test, as applied in the breadth of the present decision, incompatible with the obligations of the United States under Article 27 of TRIPS?

In relation to the first question, the brief argued that the words "directed to" lack clear definition, which may account for the risk of inappropriate paraphrasing of specific and detailed claim language leading to unacceptable flexibility.  Arbitrary disregard of claimed elements when conducting the § 101 analysis creates uncertainty for patent applicants based both in the U.S. and in Europe, for those involved in re-examination and contentious proceedings before the USPTO and for those involved in litigation.  The solution is to reaffirm the rule accepted as black letter law that all elements or limitations recited in a claim must be taken into account when considering anticipation or infringement, and hence by implication § 101 eligibility (citing the case law listed in MPEP 2131 Anticipation), and that the arrangement of those elements specified in the claim must also be considered.

Evidence of invention pertinent to § 103 should be equally applicable to § 101, and there was no evidence that the prior art paper strips nor the coding systems of Samuel Morse or Paul Revere can be said to beneficially improve the transmission of images.

Finally, it argued that the present decision is in conflict with established legal principles under the EPC as exemplified in the landmark decision T 208/84 Computer-related invention/VICOM.  In that decision, the EPO held that a claim to digitally filtering a data array was an ineligible mathematical method, but it became eligible when explicitly applied to digitally processing images.  In contrast, the panel opinion here rejects operations performed on coded data as ineligible, but refuses to acknowledge that such operations should become eligible when applied to the production of images.

It is to be hoped that the petition will be granted and that the present case will provide a landmark decision leading to a fairer and less arbitrary approach to eligibility both for computer-related inventions and for those such as Sequenom in life sciences.

* Mr. Cole is a European Patent Attorney and Partner with Lucas & Co. in Warlingham, Surrey, UK and Visiting Professor at the Centre for Intellectual Property Policy & Management at Bournemouth University.

Comments

[Martin H Snyder]

"coded data as ineligible" "such operations should become eligible when applied to the production of images"

How do you split that particular hair? How are images not data? How is the abstract idea of sampling not stone obvious in ANY context?

All you are proposing is more arbitrary line-drawing as to what is "technical" in character.

This is an information patent, and information is 100% abstract, all the time. I'm against patents on information, but if we have to have them (and I can see a rational case for SOME information patents) than we need a clear bright line as to what is too abstract for patenting. There is one; an intuitive, repeatable, easy test. If the information is being consumed by a non-human, it's an eligible process. If the value of the information is found in its meaning to human beings, not an eligible process. In this case, machines are clearly the consumers of the encoding and, the method should be eligible on THAT basis.

But there should be a second basis to eligibility beyond adherence to a statutory category. If there is no no invention, obviousness cannot be evaluated. Without an invention, there is no PHOSITA. Only an invention can be found to be obvious or not obvious or anticipated.

Ideas that don't rise to the level of inventions are not inventions, and cannot be evaluated as inventions. If a threshold inquiry determines that there is simply nothing added by the inventor other than stating a desirable outcome or assembling well-known means to an easily foreseeable end, there is not even an invention to analyze, and the idea(s) are ineligible.

In this case, sampling is ancient and the benefits of sampling entirely foreseeable. if the algo has some novelty, there could be a patent there....

Otherwise, yea, how can we sustain MPEG patents at all?

[Paul Cole]

@ Martin H Snyder

An image is a representation of a person or thing which is perceptible by a person in reflection or in emission.

Data is material in computer memory.

The skilled reader is perfectly aware of the difference. You do not need an Oxford or Harvard degree to know that.

Come to think of it there is a simple way to show the difference in the present context.

Suppose that you have burgled my house and I saw you leave. I could visit the local police station and using the FACES program there is a good chance that a likeness could be created.

A string of code data for the selected facial features could then be transmitted to a policeman having a tablet also loaded with FACES.

Nothing humanly recognizable would be produced by viewing the string of data. The policeman would be helpless to recognize you.

But if the data string were decoded and recognized at the tablet, the policeman would have a good likeness of you, would have a good chance of finding and arresting you, and you would then go to prison as richly deserved.

Your statement that this is an information patent and information is 100% abstract is what we in the UK refer to as bollocks. It is astonishing that a person of your background and knowledge should lend your name to such nonsense, even in this age of fake fact.

[skeptical]

Mr. Snyder has ever refused to recognize that ALL utility, and especially utility in the legal world of patents, touches upon the utility "to meaning to human beings."

Try the converse: what out there that has NO meaning to human beings none-the-less has patentable utility?