Unsupervised defect segmentation
An inspection system may receive inspection datasets from a defect inspection system associated with inspection of one or more samples, where an inspection dataset of the plurality of inspection datasets associated with a defect includes values of two or more signal attributes and values of one or more context attributes. An inspection system may further label each of the inspection datasets with a class label based on respective positions of each of the inspection datasets in a signal space defined by the two or more signal attributes, where each class label corresponds to a region of the signal space. An inspection system may further segment the inspection datasets into two or more defect groups by training a classifier with the values of the context attributes and corresponding class labels for the inspection datasets, where the two or more defect groups are identified based on the trained classifier.
1. An inspection system comprising:
a controller including one or more processors configured to execute program instructions configured to cause the one or more processors to:
receive a plurality of inspection datasets from a defect inspection system associated with inspection of one or more samples, wherein an inspection dataset of the plurality of inspection datasets associated with a defect includes values of two or more signal attributes and values of one or more context attributes, wherein a signal attribute of the two or more signal attributes is representative of a signal generated by the defect inspection system in response to a defect, wherein a context attribute of the one or more context attributes is representative of one or more additional characteristics of the defect;
label each of the inspection datasets with a class label from a set of class labels based on respective positions of each of the inspection datasets in a signal space defined by the two or more signal attributes, wherein each class label corresponds to a region of the signal space, wherein the one or more processors label each of the inspection datasets with a class label by: mapping the plurality of inspection datasets in the signal space; dividing the signal space into a set of polygons, wherein the inspection datasets are distributed between at least two polygons of the set of polygons; and labeling each inspection dataset of the plurality of inspection datasets with a class label from the set of class labels corresponding to a polygon from the set of polygons containing the respective inspection dataset; and
segment the plurality of inspection datasets into two or more defect groups by training a classifier with the values of the one or more context attributes and corresponding class labels from the set of class labels for the plurality of inspection datasets, wherein the two or more defect groups are identified based on the trained classifier.
2. The inspection system of claim 1 , wherein training the classifier with the values of the one or more context attributes and corresponding class labels from the set of class labels for the plurality of inspection datasets comprises:
training the classifier with the values of the one or more context attributes and corresponding class labels from the set of class labels for the plurality of inspection datasets using a supervised learning technique, wherein identifying the two or more defect groups from the trained classifier comprises:
identifying the two or more defect groups based on at least one of a learned function or a structure of the trained classifier.
3. The inspection system of claim 2 , wherein the supervised learning technique comprises:
a machine learning technique.
4. The inspection system of claim 2 , wherein the supervised learning technique comprises:
at least one of a decision tree technique, nearest neighbor technique, a regression technique, a Bayesian technique, a support vector machine technique, a similarity technique, or a neural network technique.
5. The inspection system of claim 1 , wherein the set of class labels correspond to nonoverlapping regions of the signal space.
6. The inspection system of claim 1 , wherein the set of class labels comprises:
a one-dimensional representation of the signal space defined by the two or more signal attributes.
7. The inspection system of claim 1 , wherein labeling each of the inspection datasets with a class label from the set of class labels based on respective positions of each of the inspection datasets in a signal space defined by the two or more signal attributes comprises:
generating the set of class labels corresponding to nonoverlapping regions of a multi-dimensional signal space defined by the two or more signal attributes.
8. The inspection system of claim 1 , wherein dividing the signal space into a set of polygons comprises:
dividing the signal space into the set of polygons by dividing the signal space into a regular multi-dimensional grid.
9. The inspection system of claim 1 , wherein dividing the signal space into a set of polygons comprises:
dividing the signal space into the set of polygons using at least one of Delaunay triangulation or tessellation.
10. The inspection system of claim 1 , wherein the two or more defect groups correspond to one or more contiguous groups of polygons in the set of polygons associated with the signal space defined by the two or more signal attributes.
11. The inspection system of claim 9 , wherein the one or more processors are further configured to execute program instructions causing the one or more processors to:
extract ranges of the two or more signal attributes associated with at least some of the two or more defect groups.
12. The inspection system of claim 1 , wherein the one or more processors are further configured to execute program instructions causing the one or more processors to:
classify one or more defects based on the two or more defect groups.
13. The inspection system of claim 1 , wherein the one or more processors are further configured to execute program instructions causing the one or more processors to:
generate an inspection recipe based on at least one of two or more defect groups.
14. The inspection system of claim 13 , wherein the defect inspection recipe provides detection of selected defects of interest.
15. The inspection system of claim 13 , wherein the defect inspection recipe provides nuisance filtering.
16. The inspection system of claim 1 , wherein the one or more processors are further configured to execute program instructions causing the one or more processors to:
select an optics mode based on the two or more defect groups.
17. The inspection system of claim 16 , wherein selecting the optics mode based on the two or more defect groups comprises:
ranking two or more optics modes based on one or more separation distances of one or more selected defects of interest from additional defects; and
selecting the optics mode to maximize the one or more separation distances.
18. An inspection system comprising:
a defect inspection system comprising;
an illumination source configured to generate one or more illumination beams; and
a detector configured to capture radiation emanating from a sample in response to the one or more illumination beams; and
a controller communicatively coupled to the detector, the controller including one or more processors configured to execute program instructions configured to cause the one or more processors to:
receive a plurality of inspection datasets from a defect inspection system associated with inspection of one or more samples, wherein an inspection dataset of the plurality of inspection datasets associated with a defect includes values of two or more signal attributes and values of one or more context attributes, wherein a signal attribute of the two or more signal attributes is representative of a signal generated by the defect inspection system in response to a defect, wherein a context attribute of the one or more context attributes is representative of one or more additional characteristics of the defect;
label each of the inspection datasets with a class label from a set of class labels based on respective positions of each of the inspection datasets in a signal space defined by the two or more signal attributes, wherein each class label corresponds to a region of the signal space, wherein the one or more processors label each of the inspection datasets with a class label by: mapping the plurality of inspection datasets in the signal space; dividing the signal space into a set of polygons, wherein the inspection datasets are distributed between at least two polygons of the set of polygons; and labeling each inspection dataset of the plurality of inspection datasets with a class label from the set of class labels corresponding to a polygon from the set of polygons containing the respective inspection dataset; and
segment the plurality of inspection datasets into two or more defect groups by training a classifier with the values of the one or more context attributes and corresponding class labels from the set of class labels for the plurality of inspection datasets, wherein the two or more defect groups are identified based on the trained classifier.
19. The inspection system of claim 18 , wherein the illumination source comprises:
an optical illumination source.
20. The inspection system of claim 19 , wherein the illumination source comprises:
a broadband plasma illumination source.
21. The inspection system of claim 19 , wherein wavelengths of the one or more illumination beams comprise:
at least one of extreme ultraviolet wavelengths, vacuum ultraviolet wavelengths, deep ultraviolet wavelengths, or ultraviolet wavelengths.
22. The inspection system of claim 18 , wherein the illumination source comprises:
a particle-beam illumination source.
23. The inspection system of claim 22 , wherein the illumination source comprises:
an electron-beam illumination source.
24. The inspection system of claim 18 , wherein training the classifier with the values of the one or more context attributes and corresponding class labels from the set of class labels for the plurality of inspection datasets comprises:
training the classifier with the values of the one or more context attributes and corresponding class labels from the set of class labels for the plurality of inspection datasets using a supervised learning technique, wherein identifying the two or more defect groups from the trained classifier comprises:
identifying the two or more defect groups based on at least one of a learned function or a structure of the trained classifier.
25. The inspection system of claim 24 , wherein the supervised learning technique comprises:
a machine learning technique.
26. The inspection system of claim 24 , wherein the supervised learning technique comprises:
at least one of a decision tree technique, nearest neighbor technique, a regression technique, a Bayesian technique, a support vector machine technique, a similarity technique, or a neural network technique.
27. The inspection system of claim 18 , wherein the set of class labels correspond to nonoverlapping regions of the signal space.
28. The inspection system of claim 18 , wherein labeling each of the inspection datasets with a class label from the set of class labels based on respective positions of each of the inspection datasets in a signal space defined by the two or more signal attributes comprises:
generating the set of class labels corresponding to nonoverlapping regions of a multi-dimensional signal space defined by the two or more signal attributes.
29. The inspection system of claim 18 , wherein dividing the signal space into a set of polygons comprises:
dividing the signal space into the set of polygons by dividing the signal space using at least one of a multi-dimensional grid, Delaunay triangulation or tessellation.
30. The inspection system of claim 18 , wherein the one or more processors are further configured to execute program instructions causing the one or more processors to:
extract ranges of the two or more signal attributes associated with at least some of the two or more defect groups.
31. The inspection system of claim 18 , wherein the one or more processors are further configured to execute program instructions causing the one or more processors to:
classify one or more defects based on the two or more defect groups.
32. The inspection system of claim 18 , wherein the one or more processors are further configured to execute program instructions causing the one or more processors to:
generate an inspection recipe based on at least one of two or more defect groups.
33. The inspection system of claim 32 , wherein the defect inspection recipe provides nuisance filtering.
34. The inspection system of claim 18 , wherein the one or more processors are further configured to execute program instructions causing the one or more processors to:
select an optics mode based on the two or more defect groups.
35. A method for segmenting metrology defects comprising:
receiving a plurality of inspection datasets from a defect inspection system associated with inspection of one or more samples, wherein an inspection dataset of the plurality of inspection datasets associated with a defect includes values of two or more signal attributes and values of one or more context attributes, wherein a signal attribute of the two or more signal attributes is representative of a signal generated by the defect inspection system in response to a defect, wherein a context attribute of the one or more context attributes is representative of one or more additional characteristics of the defect;
labeling each of the inspection datasets with a class label from a set of class labels based on respective positions of each of the inspection datasets in a signal space defined by the two or more signal attributes, wherein each class label corresponds to a region of the signal space, wherein the labeling each of the inspection datasets with a class label comprises mapping the plurality of inspection datasets in the signal space; dividing the signal space into a set of polygons, wherein the inspection datasets are distributed between at least two polygons of the set of polygons; and labeling each inspection dataset of the plurality of inspection datasets with a class label from the set of class labels corresponding to a polygon from the set of polygons containing the respective inspection dataset; and
segmenting the plurality of inspection datasets into two or more defect groups by training a classifier with the values of the one or more context attributes and corresponding class labels from the set of class labels for the plurality of inspection datasets, wherein the two or more defect groups are identified based on the trained classifier.
36. An inspection system comprising:
a controller including one or more processors configured to execute program instructions configured to cause the one or more processors to:
receive a plurality of inspection datasets from a defect inspection system associated with inspection of one or more samples, wherein an inspection dataset of the plurality of inspection datasets associated with a defect includes values of two or more signal attributes and values of one or more context attributes, wherein a signal attribute of the two or more signal attributes is representative of a signal generated by the defect inspection system in response to a defect, wherein a context attribute of the one or more context attributes is representative of one or more additional characteristics of the defect;
label each of the inspection datasets with a class label from a set of class labels based on respective positions of each of the inspection datasets in a signal space defined by the two or more signal attributes, wherein each class label corresponds to a region of the signal space;
segment the plurality of inspection datasets into two or more defect groups by training a classifier with the values of the one or more context attributes and corresponding class labels from the set of class labels for the plurality of inspection datasets, wherein the two or more defect groups are identified based on the trained classifier; and
select an optics mode based on the two or more defect groups by ranking two or more optics modes based on one or more separation distances of one or more selected defects of interest from one or more additional defects; and selecting the optics mode to maximize the one or more separation distances.