DETAILED ACTION
Notice of Pre-AIA or AIA Status
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
Drawings
Figures 4 and 5A should be designated by a legend such as --Prior Art-- because only that which is old is illustrated. See MPEP § 608.02(g). Corrected drawings in compliance with 37 CFR 1.121(d) are required in reply to the Office action to avoid abandonment of the application. The replacement sheet(s) should be labeled “Replacement Sheet” in the page header (as per 37 CFR 1.84(c)) so as not to obstruct any portion of the drawing figures. If the changes are not accepted by the examiner, the applicant will be notified and informed of any required corrective action in the next Office action. The objection to the drawings will not be held in abeyance.
Claim Rejections - 35 USC § 112
The following is a quotation of 35 U.S.C. 112(b):
(b) CONCLUSION.—The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the inventor or a joint inventor regards as the invention.
The following is a quotation of 35 U.S.C. 112 (pre-AIA ), second paragraph:
The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the applicant regards as his invention.
Claims 1-20 are rejected under 35 U.S.C. 112(b) or 35 U.S.C. 112 (pre-AIA ), second paragraph, as being indefinite for failing to particularly point out and distinctly claim the subject matter which the inventor or a joint inventor (or for applications subject to pre-AIA 35 U.S.C. 112, the applicant), regards as the invention.
Claim 1 is rejected for the following reasons:
i) The limitation “first classification phase” renders the claim indefinite. It is unclear and confusing what is referred to by a “phase”. For example, a well-known learning algorithm would go through an initial training for adjusting parameters followed by testing and often re-training if necessary. Do each of these steps alone or together constitute a phase? Please provide clarification by pointing to the portion of the applicant’s specification that defines the claimed “phase” or “first classification phase”, or amend the claim for clarification.
ii) Consequently, the limitation “multi-phase classification” renders the claim indefinite. It is unclear and confusing what is referred to by “multi-phase”. The applicant points to systems 400 and 500 of fig 4 and 5A as depicting a single-phase classification, and points to system 600 in fig 6 as multi-phase classification. The only difference between the multi-phase and single-phase classification appears to be whether re-training is provided or not. Please confirm so, or amend the claim for clarification.
Similar reasons apply to claims 8 and 14 that recite subject matter similar to claim 1.
Claim 2 is rejected for the following reasons:
iii) The limitation “performing a second classification phase” renders the claim indefinite. It is unclear and confusing whether such second classification phase is part of or separate from the “multi-phase classification” of claim 1. It is also unclear and confusing when such “second classification phase” is performed with respect to the steps recited in claim 1. For example, is the second classification phase performed prior to or after applying the multi-phase classification? Please amend the claim for clarification.
Similar reasons apply to claims 9 and 15 that recite subject matter similar to claim 2.
Claim 3 is rejected for the following reasons:
iv) The limitation “based on a defect review type” renders the claim indefinite. It is unclear and confusing whether the defect review type is included in the “plurality of defect review types” recited in claim 1. If so, please amend the claim to recite “based on a defect review type of the plurality of defect review types” for clarification and maintaining consistency in claim language.
Claim 4 is rejected for the following reasons:
v) The limitation “data imbalance” renders the claim indefinite. It is unclear and confusing in what perspective the data is imbalanced. For example, does the mere presence of mis-classified defects introduce imbalance?
vi) Consequently, the limitation “developed to reduce a data imbalance” renders the claim indefinite. It is further unclear and confusing whether this phrase is merely a recitation of intended use or purpose of the claimed invention. For example, is the data imbalance actually quantified before and after developing the plurality of nuisance review types for comparison?
Similar reasons apply to claims 11 and 17 that recite subject matter similar to claim 4.
Claim 5 is rejected for the following reasons:
vii) The limitation “developed to yield a classification count” renders the claim indefinite. It is unclear and confusing whether this phrase is merely a recitation of intended use or purpose of the claimed invention. For example, is the classification count actually quantified before and after developing the first nuisance review type for comparison?
viii) The limitation “the classification count of the second nuisance review type” lacks antecedent basis.
Similar reasons apply to claims 12 and 18 that recite subject matter similar to claim 5.
Claim Rejections - 35 USC § 103
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
Claims 1, 4, 6-8, 11, 13, 14, 17, 19, and 20 are rejected under 35 U.S.C. 103 as being unpatentable over Plihal et al. (US 2019/0067060) in view of Chen et al. (US 10,545,099).
Regarding claim 1, Plihal discloses:
obtaining image data comprising a set of candidate defects (see para [48]-[50], obtaining image data of defect candidates);
developing a plurality of defect review types and a plurality of nuisance review types (see [126], a plurality of DOI (defects of interest) examples and a plurality of nuisance examples in a training set indicate an inherent developing of such examples);
classifying the set of candidate defects into one or more defect types based on the plurality of defect review types during a first classification phase (see [51], [115], [124], and [135], classifying some of the defect candidates as “DOIs” based on a learning algorithm trained on the plurality of DOI examples and nuisance examples);
classifying the set of candidate defects into a plurality of nuisance types based on the plurality of nuisance review types during the first classification phase (see [51], [115], [124], and [135], classifying some of the defect candidates as “nuisances” based on the learning algorithm trained on the plurality of DOI examples and nuisance examples); and
applying a machine learning based multi-phase classification (see [122], classification results of the learning algorithm are used for re-training; and see [134], the learning algorithm may have multiple layers).
However, Plihal does not disclose: to the classified set of candidate defects (i.e., Plihal discloses classifying defect candidates as either DOI or nuisance, and the usage of a re-training for a learning algorithm, however, does not disclose a secondary learning algorithm is applied to results of a first learning algorithm).
In a similar field of endeavor of distinguishing DOIs and nuisances via a learning algorithm, Chen discloses: to the classified set of candidate defects (see col 6 line 30 through col 7 line 4 and fig 2, further applying a second learning algorithm based classification in step 206, after defect candidates are classified as DOIs or nuisances by a first learning algorithm in step 204, in order to classify the classified DOIs into different types of DOIs).
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to combine Plihal with Chen, and provide a learning algorithm to classify defect candidates into DOIs and nuisances, as disclosed by Plihal, and provide an additional learning algorithm to further classify the DOIs into DOI types, as disclosed by Chen, for the purpose of achieving a semiconductor inspection with improved sensitivity to defects (see Chen 1:52-2:10).
Regarding claim 4, Plihal further discloses: wherein at least one of the plurality of nuisance review types is developed to reduce a data imbalance in the classified image data set of candidate defects (see [6] and rejection of claim 1, the developed nuisance examples are for the purpose of suppressing nuisance rates in the defect candidates).
Regarding claim 6, Plihal further discloses: wherein the set of candidate defects is from a charged particle beam apparatus including a detector (see [69], charged particle beam-based tool and/or a SEM (scanning electron microscope)).
Regarding claim 7, Plihal further discloses: wherein the charged particle beam apparatus including a detector is a scanning electron microscope (SEM) (see rejection of claim 6, SEM).
Regarding claim 8, Plihal and Chen disclose everything claimed as applied above (see rejection of claim 1). Plihal further discloses:
a charged particle beam apparatus including a detector (see [69], charged particle beam-based tool);
an image acquirer that includes circuitry to receive a detection signal from the detector and construct an image including a first feature (see [32], imaging detectors); and
a controller with at least one processor and a non-transitory computer readable medium comprising instructions that, when executed by the at least one processor, cause the system (see [34]-[35], a computer).
Regarding claims 11 and 13, Plihal and Chen disclose everything claimed as applied above (see rejection of claims 4, 7, and 8).
Regarding claims 14, 17, 19, and 20, Plihal and Chen disclose everything claimed as applied above (see rejection of claims 1, 4, 6, and 7).
Claims 2, 9, and 15 are rejected under 35 U.S.C. 103 as being unpatentable over Plihal and Chen in view of Cahoon et al. (US 11,880,746).
Regarding claim 2, Plihal and Chen disclose everything claimed as applied above (see rejection of claim 1), however, do not disclose: wherein the machine learning based multi-phase classification further comprises: performing a review of the classified set of candidate defects from the first classification phase; selecting at least one of a misclassified defect and a misclassified nuisance from the classified set of candidate defects; re-labeling the at least one of the misclassified defect and the misclassified nuisance according to the review to create relabeled image data comprising the set of candidate defects; adding the relabeled image data to a training pool of the machine learning based multi-phase classification to create a revised training pool; and performing a second classification phase using the revised training pool (i.e., Plihal discloses, in [122], that the classification results of the learning algorithm are used for re-training the learning algorithm, however, does not specify that it includes re-labeling misclassified results for re-training).
In a similar field of endeavor of utilizing a neural network for a classification task, Cahoon discloses: wherein the machine learning based multi-phase classification further comprises: performing a review of the classified set of candidate defects from the first classification phase; selecting at least one of a misclassified defect and a misclassified nuisance from the classified set of candidate defects; re-labeling the at least one of the misclassified defect and the misclassified nuisance according to the review to create relabeled image data comprising the set of candidate defects; adding the relabeled image data to a training pool of the machine learning based multi-phase classification to create a revised training pool; and performing a second classification phase using the revised training pool (see col 5 lines 3-18, reviewing a classification result of a learning algorithm to select misclassified images, relabeling the misclassified images, and further training the learning algorithm by adding the relabeled misclassified images into the training set).
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to combine Plihal and Chen with Cahoon, and provide a retrainable learning algorithm trained on training data of DOI examples and nuisance examples to classify a defect image, as disclosed by Plihal and Chen, and further provide an expert to manually relabel misclassified images for retraining the learning algorithm, as disclosed by Cahoon, for the purpose of achieving accurate classification (see Cahoon 5:3-18).
Regarding claims 9 and 15, Plihal, Chen, and Cahoon disclose everything claimed as applied above (see rejection of claims 2, 8, and 14).
Claims 3, 10, and 16 are rejected under 35 U.S.C. 103 as being unpatentable over Plihal and Chen in view of Brauer (US 2019/0073566).
Regarding claim 3, Plihal and Chen disclose everything claimed as applied above (see rejection of claim 1), however, do not disclose: wherein at least one of the plurality of nuisance review types is developed based on a defect review type (i.e., Plihal discloses developing a plurality of DOI examples and nuisance examples, however, does not disclose that the nuisance examples are based on the DOI examples).
In a similar field of endeavor of distinguishing DOIs and nuisances via a learning algorithm, Brause discloses: wherein at least one of the plurality of nuisance review types is developed based on a defect review type (see [75], the number of nuisance examples are set based on the number of DOI examples in the training set).
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to combine Plihal and Chen with Brauer, and provide a training set of DOI examples and nuisance examples, as disclosed by Plihal and Chen, wherein the number of nuisance examples are set based on the number of DOI examples, as disclosed by Brauer, for the purpose of neutralizing bias provided by the training set (see Brauer [75]).
Regarding claims 10 and 16, Plihal, Chen, and Brauer disclose everything claimed as applied above (see rejection of claims 3, 8, and 14).
Allowable Subject Matter
The prior art of record does not disclose the subject matter recited in claims 5, 12, and 18, however, these claims are rejected under 112(b). These claims would be allowable if amended to overcome the 112(b) rejection and rewritten in independent form including all of the limitations of the base claim and any intervening claims. The following is a statement of reasons for the indication of allowable subject matter:
Regarding claim 5, Plihal and Chen do not disclose: wherein the at least one of the plurality of nuisance review types includes a first nuisance review type and a second nuisance review type; and the first nuisance review type is developed to yield a classification count that is not more than 5 times the classification count of the second nuisance review type. Similar reasons apply to claims 12 and 18.
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure
Shaubi et al. (US 2019/0066290) discloses utilizing a neural network to distinguish defects vs nuisances.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to SJ PARK whose telephone number is (571)270-3569. The examiner can normally be reached M-F 8:00 AM - 5:00 PM.
Examiner interviews are available via telephone, in-person, and video conferencing using a USPTO supplied web-based collaboration tool. To schedule an interview, applicant is encouraged to use the USPTO Automated Interview Request (AIR) at http://www.uspto.gov/interviewpractice.
If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, EMILY TERRELL can be reached at 571-270-3717. The fax phone number for the organization where this application or proceeding is assigned is 571-273-8300.
Information regarding the status of published or unpublished applications may be obtained from Patent Center. Unpublished application information in Patent Center is available to registered users. To file and manage patent submissions in Patent Center, visit: https://patentcenter.uspto.gov. Visit https://www.uspto.gov/patents/apply/patent-center for more information about Patent Center and https://www.uspto.gov/patents/docx for information about filing in DOCX format. For additional questions, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free). If you would like assistance from a USPTO Customer Service Representative, call 800-786-9199 (IN USA OR CANADA) or 571-272-1000.
/SJ Park/Primary Examiner, Art Unit 2675
Prosecution Insights