T 1997/11 - Common general knowledge

Key points

• The Board considers the invention to be insufficiently disclosed, because D0,  a more detailed publication by some of the inventors (with a publication date after the priority date) on how the present invention is implemented refers to D3 (an earlier published article) for critical details of the used algorithm. The application did not refer to D3 and D3 is, as scientific article, not common general knowledge.
  
  
  EPO T 1997/11 - link

Reasons for the Decision

1. The appeal is admissible.

2. Sufficiency of disclosure (Article 83 EPC)

2.1 The appellant argued in the statement of grounds of appeal that the skilled person could have implemented the invention given the application description and having D3 available as part of his common general knowledge.

2.2 Hence, the first question to answer is whether D3 forms part of the common general knowledge of the skilled person in the field of blind deconvolution. The Board considers that this is not the case for the following reasons.

2.2.1 Document D3 is a scientific article referring to blind deconvolution of ultrasonic pulses for non-destructive evaluation of an object. As indicated in the decision under appeal with reference to established jurisprudence (Case Law of the Boards of Appeal, 6th edition 2010, I.C.1.5), the common general knowledge is normally represented by encyclopaedias, textbooks, dictionaries and handbooks. Only by way of exception, patent specifications and scientific publications have been considered to form part of the common general knowledge, e.g. if the field of research was so new that technical knowledge was not yet available from textbooks. Blind deconvolution is however by no means a new field of research, a fact which is not in dispute. Indeed, the appellant during the first-instance proceedings cited numerous other scientific publications dealing with blind deconvolution algorithms (referred to as D15 to D24), and a monograph was filed with the statement of grounds of appeal likewise showing that blind deconvolution is well known.

2.2.2 Among the wealth of publications dealing with blind deconvolution algorithms, D3 may be relevant merely in order to explain certain specific aspects of the algorithm with which the present application is concerned. In fact, D3 could only have been found as a result of a comprehensive search among the many specialised scientific publications in the field.

2.2.3 The Board therefore concludes that D3 does not reflect the common general knowledge of the skilled person in the field of blind deconvolution.

2.3 In these circumstances, the next question to answer is whether the skilled person could have implemented the invention on the basis of the application alone using common general knowledge in the field of blind deconvolution.

2.3.1 In the claimed invention (of all requests), a final set of ultrasound response data for the vascular tissue is calculated from (at least) the estimated first set of ultrasound response data (Xest), the estimated transfer function, and the (received) ultrasound data (Z).

2.3.2 The description of the application fails to describe at least one example of how this calculation is carried out, as required by Rule 42(1)(e) EPC.

The description dedicates just three lines (on page 8, lines 7 to 10) to describing how the estimate of X (Xest) and the estimate of the transfer function H is obtained. It indicates that the algorithm of the invention searches first for an estimate of X (Xest) that minimizes a certain error criteria. This criteria measures the differences between the observed data (Z) and the fitted model. Then H is estimated by a least-squares-fit algorithm.

2.3.3 It is pertinent to compare the present application to D0, an even more detailed publication by some of the inventors (with a publication date after the priority date) on how the present invention is implemented. Therein the authors explain (in the two first paragraphs of page 1128) that H is obtained in an iterative maximization estimation, and they refer the reader to citation [5] - which is document D3 - to obtain there the necessary details of the algorithm. D0 mentions a further particularly relevant aspect of the algorithm, i.e. the introduction of a cost function incorporating a penalty term which favours solutions with as few non-zero data Xi as possible, hence ensuring sparse reflectivity as occurs in intravascular ultrasound backscatter measurements (D3, page 564, right column, first portion of the third paragraph). Moreover, the appellant highlighted in the statement of grounds of appeal (point 4.3) the fact that D3 stresses that "sparsity of the reflectivity is fundamental to the approach" (page 572, last full paragraph of the left column; abstract; page 564, right column, second full paragraph; page 566, left column, last paragraph). Hence, according to D3, starting from the last paragraph on page 566, standard sparse matrix techniques are to be used. In section C of D3 these are applied to obtain the iterative estimate of Hi (hi in D3).

It is therefore clear that without the reference to D3 in D0 the reader of D0 would not have been able to reproduce the blind deconvolution algorithm of D0.

Likewise, since the application omits all the detailed and complex aspects of the algorithm described in D0 in combination with D3, the skilled person would not have been in a position to carry out the invention using just standard mathematical skills.

2.4 As a consequence, the Board finds that the application does not disclose the invention in a manner sufficiently complete for it to be carried out by a person skilled in the art, as required by Article 83 EPC.

Order

For these reasons it is decided that:

The appeal is dismissed.