The European Patent Office refused to grant a patent on a method of authenticating financial transactions based on biometric data. Here are the practical takeaways of the decision T 1386/14 (Fraud resistant biometric financial transaction / EYELOCK LLC) of 12.6.2019 of Technical Board of Appeal 3.4.01:
This European patent application concerns biometric identification and authentication methods, in particular for authenticating financial transactions using biometrics.
According to the application, one problem faced by biometric recognition systems involves the possibility of spoofing, e.g. when an imposter presents a life-sized, high-resolution photograph of a person to an iris recognition system. To this end, known systems try to determine liveness of the subject to be authenticated, e.g. by shining a light onto the eye and determining whether the pupil dilates. The patent application criticizes that in prior art systems which determine liveness, the liveness test is conducted first, prior to the match process, i.e. the calculation of the probability of a match between acquired biometric data from the individual being authenticated and data acquired from known individuals.
Here is how the invention is defined in claim 1 (auxiliary request 4):
Claim 1 (auxiliary request 4)A method of authorising financial transactions comprising:
– acquiring biometric data from a person;
– calculating a probability of liveness, Pp, being a probability that biometric data has been acquired that can be used to identify the person after the fact;
– calculating a probability of a match, Pm, being a probability of matching between the person and known persons whose biometric information has previously been acquired;
– authorising or not authorising said transaction based on a combination of probabilities Pp and Pm in such a manner that transactions will be authorised when the probability of liveness is high even if the probability of a match is poor, and wherein in the event of non authorisation the person is not informed whether non-authorisation was based on a low probability of liveness or a low probability of matching.
Is it technical?
The board of appeal based its assessment of inventive step on a prior art document which concerned biometric authentication systems:
Document D8 concerns biometric authentication techniques. It focuses on the merits of multibiometric systems, that is, on systems making use of multiple biometric sensors for data acquisition. (…) In one category of multibiometric systems, data are fused together at the decision level, that is, at the end of the decision process. Such a configuration, as described in section “2.2 Fusion at the Matching Score Level” of D8, was considered as closest prior art by the Examining Division in its rejection of auxiliary request 2 (then pending) for lack of inventive step. (…) In view of the fact that multibiometric systems indeed address the problems of reliability and accuracy of matching and further attempt to remedy the problems of spoofing identified with regard to previous techniques (cf. D8, section “Introduction”, first paragraph), the selection of D8 as starting point is justified.
First of all, the board criticized the first instance examining division for failing to apply the proper problem-solution approach:
Although details are lacking, it seems that the Examining Division meant that it would have been obvious either to replace one of the sensors of Figure 3 in D8 by a sensor indicative of “liveness” or to add such a sensor.
The Examining Division defined neither the objective technical problem solved by the claimed invention, nor the technical features distinguishing the claimed subject-matter from D8.
Both are, however, essential in the problem – solution approach. A prerequisite for the existence of an inventive step is the existence of a technical difference that provides a technical effect.
Concerning inventive step, the board took the view that the feature of calculating a probability of “liveness” is a non-technical criterion which thus cannot establish non-obviousness:
The method according to claim 1 of the fourth auxiliary request differs from the method of D8 in that it calculates a probability of “liveness”, that is, a probability that biometric data have been “acquired from a live human”. (This construes the unclear wording in claim 1, in the appellant’s favour, as reflecting the teaching of paragraph  of the published application). The choice of “liveness” as criterion is a non-technical decision as to what sort of identification is acceptable. The combination of probabilities, itself, is a mathematical operation and does not contribute to the technical character of the claimed method, at least insofar as it is construed as referring to mere calculations carried out on the available biometric data. The Board notes that the claim is not concerned with the technical means by which “liveness” is measured, but only with the fact that it is “liveness”.
The same conclusion would apply to the system of independent claim 14, which requires a corresponding calculating means for calculating the probability of “liveness”. Since the system of D8 also comprises calculating means, the contribution is limited to the calculation being carried out, as such.
A consequence of the previous analysis is that no distinction of a technical nature can be identified between the claimed inventions and the disclosure of D8 (cf. T 119/11, for example).
For these reasons, the board concluded that claim 1 does not involve an inventive step.
You can read the whole decision here: T 1386/14 (Fraud resistant biometric financial transaction / EYELOCK LLC) of 12.6.2019