Camtek Invalidates another Orbotech Patent

Duck Shoot

IL 131283 titled ” Apparatus and methods for inspection of objects” was filed on 5 August 1999. After its eventual allowance it published for opposition purposes on 8 July 2008. On 7 October 2008 Camtek filed a notice of opposition and on 5 November 2008 they filed their Statement of Case.

Both sides submitted their evidence and on 28 March 2010, the Applicant requested to correct the specification, and submitted amended claims. On 24 May 2010 Camtek announced that they had no objections to the amendments, but reserved the right to raise further issues in the opposition itself. The parties submitted their statements and evidence regarding the amended claims, and, after a hearing back in March 2013, this ruling has now issued.

The patent in question relates to automated optical inspection and the main claim of the patent in question is as follows:

A method for image processing comprising:

generating representations of boundaries of known elements in the image;

identifying candidate defects in the image, in hardware, thereby to generate a hardware candidate defect identifier output including hardware candidate defect identifiers;

receiving the hardware candidate defect identifiers and said representations of boundaries;

generating a window of  interest surrounding at least some of said hardware candidate defect identifiers and said representations of boundaries;

identifying at least one false alarm within said window of interest, in software;

and identifying additional candidate defects in the image, in software.

There are a further 6 dependent claims.

The underlined parts were the result of the amendment during the opposition, which, since it is a narrowing restriction, is procedurally acceptable.

Opposer’s Case

In the Statement of Opposition, Camtek opposed the patent issuing on the basis of Sections 3, 4, 5, 12, 13 and 18 of the Law.

Re Section 3 – Camtek alleged that the patent isn’t directed to a device or process, and at best is an attempt to monopolize a computer process or algorithm per se.

Camtek considers the claimed invention to lack novelty as anticipated separately in three earlier publications:

  1. West at al, ‘The Automatic Visual Inspection of Printed Circuit Boards’ Circuit World Vo. 8 No.2 (1982)
  2. West, ‘A system for the Automatic Visual Inspection of Bare-Printed Circuit Boards’, IEER Transactions on Systems, Man and Cybernetics, Vol. SMC -14, No.5 (1984)
  3. US 5,619,429; M. Aloni at al, Apr. 8 1997

These publications allegedly render the claimed invention obvious, whether considered separately or in combination.

Camtek further accused Orbotech of inequitable behavior in having not brought the existence of US 5,619,429 (which was assigned to Orbot Instruments) to the attention of the Examiner contrary to the duty of disclosure. Camtek also alleged that the amendments to claim 1 designed to attempt to distinguish the claimed invention from that of 5,619,429 did nothing to rectify the failure of Orbotech to bring this to the Examiner’s attention. Finally, Camtek argued that the claims lacked clarity and were not supported by the specification.

Orbotech’s Position

Orbotech denies all of Camtek’s allegations. According to Orbotech, the claimed invention differs from each of the cited art by including at least one element not described therein. In particular, none of the citations relates the step of:

“Generating a window of interest surrounding at least some of said hardware candidate defect identifiers and said representations of boundaries”

Orbotech did not consider it is required to address the issue of inventive step (obviousness) since Camtek had merely submitted a list of prior art references and had not shown how they can be combined to teach the claimed invention. Similarly, the lack of clarity and support issues were denied and Orbotech noted that they weren’t substantiated in Camtek’s statement of case.

Evidence

Camtek’s position was supported by an affidavit from Dr Shimon Koren, the director of Technology in the Microelectronics Division at Camtek. Koren considered that the claimed invention is a known methodology for image analysis that combines hardware and software stages and a standard way of defining areas of interest. He considers that using software or dedicated hardware was simply a design choice and the claimed elements were the foundation of software engineering in the field, and known to all persons of the art at the priority date. Specifically, any further image processing on a feature of interest requires the preliminary determination of a boundary to the region of interest, whether automated, by hardware or software, or manually, and there was nothing novel and certainly not inventive in this. [I find this argument compelling, as I have written up about a dozen patents for Siemens CAD vision for detecting lung and breast cancer by image analysis of XRay photographs, and everything starts with drawing a line around the field of interest].

Koren does not understand the relevance of “identifying at least one false alarm within said window of interest, in software; and identifying additional candidate defects in the image, in software.” And, as a person of the art, Korens considers that these claimed steps are simply confusing, hence the lack of clarity objection. Koren further expounded on the lack of novelty and inventiveness with respect to the prior art referenced above.

Orbotech submitted an Affidavit of the Director of their Core Technology Algorithm Group, a Dr Rippa, who is a coinventor of the claimed invention.

Dr Rippa stated that the present invention claims a process for identifying and segmenting candidate defects or preselected areas of interest, whereas the prior art merely describes a process for identifying and segmenting candidate defects. Dr Rippa concedes that the hierarchical examinations where each stage filters through candidate defects picked up by the proceeding stage were known to persons of the art. However, the present invention is not necessarily hierarchical in that the areas of interest with their generated boundaries are not necessarily subjected to additional filtering steps. Furthermore, algorithms per se. are not claimed, but rather a process ran on a physical system [so the invention is to be considered as patentable subject matter].

Additionally, Dr Rippa denied the allegations of lack of clarity, claiming that repeated indications provide greater certainty. [I can’t see how this jells with previous statement about the method not necessarily being hierarchical, nor can I see how the wording of the claim supports this interpretation].

Furthermore Dr Rippa also considers that there are significant differences between hardware and software and using hardware where the claimed invention specifies software will be inefficient.

Finally, Dr Rippa related to the cited art and opined that they do not render the claimed invention known or obvious.

The Ruling

Ms Yaara Shoshani Caspi considered that the invention related to the ‘window’ which was not properly described in the application. She found Dr Koren’s explanation convincing, whereas she found Dr Rippa’s explanation incoherent. She also noted that Dr Rippa, as an inventor and witness for the defendant, was more likely to provide misleading testimony. [The ‘defendant’ is the applicant. The issue is an opposition. Both sides have financial interests or the patent would not be opposed. Camtek are not acting out of respect for the integrity of the register].

It seems that during cross-examination, Dr Rippa changed his story, and didn’t seem able to make up his mind whether candidates were segmented or whether also previously defined areas were segmented.

Camtek and their witness, Dr Koren argued that once an area is selected, it requires to be segmented and the window drawing exercise is the same regardless. The point of novelty was, in their opinion, a question of semantics and not  one of substance. The issue is not one of terminology but whether Orbotech was describing something new or not.

It seems that Dr Rippa retracted his original statement and accepted that once a window was drawn, the enclosed space was checked as to whether it fulfilled the criteria of defects or not. Dr Koren argued convincingly that the examination of predetermined windows was for calibration purposes, to help recognize defects. Since Orbotech failed to convincingly argue that hardware was advantageous over software for this purpose, the argument that hardware was chosen specifically was not given any weight. The upshot of the above is that Orbotech was claiming to segment candidate flaws and then to examine them and the adjudicator did not consider this novel or inventive. The additional step  of finding additional candidates was not considered clear and could not, therefore, be used as saving grace. In this regard, the adjudicator noted a dissonance between the claim, the representative figure and the affidavit.

Ms Shoshani-Caspi, did, however, find that the although directed to an algorithm, since the algorithm was for use on a machine having certain functionality, the invention should be considered technological and not software per se or a thought process.

Having defined the invention as segmenting candidate flaws and then examining them, the adjudicator found that each of the publications:

  • West at al, ‘The Automatic Visual Inspection of Printed Circuit Boards’ Circuit World Vo. 8 No.2 (1982)
  • West, ‘A system for the Automatic Visual Inspection of Bare-Printed Circuit Boards’, IEER Transactions on Systems, Man and Cybernetics, Vol. SMC -14, No.5 (1984)
  •  US 5,619,429; M. Aloni at al, Apr. 8 1997

Separately anticipates  the claimed invention, rendering it known in contradistinction to Section 4 of the Patent Law 1967. Being known, the claimed invention could not be considered inventive in contradistinction to Section 5 of the Patent Law 1967.

That said, the adjudicator did not find the claims unsupported or evidence of inequitable behavior.

Despite the Opposer failing to specify the flaws of the dependent claims, the adjudicator did not find redeeming features in any of these. The patent was found to be invalidated and costs of 15000 Shekels were awarded to Camtek.

Comment

Once again, Camtek, represented by Adv.Adi Levit, has successfully invalidated an Orbotech patent. However, unlike the recent case where Camtek invalidated Orbotech’s IL 179995 patent and the Deputy Commissioner Ms Jacqueline Bracha ruled costs of 296,895 Shekels.as Orbotech forgot to appeal Camtek’s detailed costs submission, the costs awarded in this case were a mere 15000 Shekels which does not seem to be a lot for a case that has been dragging on for 7 years.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: