Power assisted automatic supervised classifier creation tool for semiconductor defects

• US 6,999,614 B1
• Filed: 11/28/2000
• Issued: 02/14/2006
• Est. Priority Date: 11/29/1999
• Status: Active Grant

• Current Assignee: KLA-Tencor Corporation
• Original Assignee: KLA-Tencor Corporation
• Inventors: Banerjee, Saibal, Bakker, David, Smith, Ian R.
• Primary Examiner(s): Mehta, Bhavesh M.
• Assistant Examiner(s): CARTER, AARON W
• Application Number:
  US09/724,633
• Time in Patent Office: 1,904 Days
• Field of Search: 382141-152, 382155-161, 382/224
• US Class Current: 382/159
• CPC Class Codes: G06T 2207/10056 Microscopic image
• CPC Class Codes: G06T 2207/20081 Training; Learning
• CPC Class Codes: G06T 2207/20092 Interactive image processing based on input by user
• CPC Class Codes: G06T 2207/30148 Semiconductor; IC; Wafer
• CPC Class Codes: G06T 7/0004 Industrial image inspection

Abstract

A method and system that optionally allows a user to view image defects organized by natural groupings based on features of the images. The natural groupings make it easier for the user to organize some or all of the images into classes in a training set of images. A feature vector is extracted from each image in the training set and stored, along with its user-specified class, for use by an automatic classifier software module. The automatic classifier uses the stored feature vectors and classes to automatically classify images not in the training set. If the automatically classified images do not match images manually classified by the user, the user modifies the training set until a better result is obtained from the automatic classifier. The system can provide feedback to an inspection system designed to aid in the setup and fine-tuning of the inspection system.

Claims

1. A method for classifying a plurality of images, comprising:

providing a working set of images;

prior to a user performing any classification of the working set of images, automatically sorting the working set of images into a plurality of groupings based on common features of the working set of images and displaying such groupings; and

after automatically sorting the working set of images into groupings, receiving input from the user to manually classify at least a subset of the working set of images facilitated by the displayed groupings.

2. The method of claim 1, wherein the groupings are initially displayed as a plurality of elements, wherein each element specifies a number that indicates how many of the working set of images are grouped together.

3. The method of claim 2, wherein each element is selectable by the user to thereby display the corresponding one or more working set of images that are grouped together.

4. The method of claim 1, further comprising:

providing a training set, wherein the training set is formed from the user'"'"'s manually classified subset of the working set of images; and

automatically classifying the unclassified working set of images based on a plurality of features extracted from the training set and the working set of images and the user'"'"'s manual classification of the training set.

5. The method of claim 4, further comprising displaying a visual representation of a comparison between the automatic classification and the manual classification performed by the user.

6. The method of claim 5, further comprising receiving input from the user to alter the training set based on the displayed visual representation of the comparison between the automatic classification and the manual classification performed by the user so that the automatic classification more closely matches the manual classification.

7. The method of claim 5, further comprising receiving input from the user to alter one or more parameters of the automatic classification based on the displayed visual representation of the comparison between the automatic classification and the manual classification performed by the user so that the automatic classification more closely matches the manual classification.

8. The method of claim 7, wherein altering the one or more parameters of the automatic classification includes graphically manipulating one or more images.

9. The method of claim 4, further comprising automatically classifying a second working set of images based on a plurality of features extracted from the training set and the second working set of images and the user selected classes of the training set.

10. The method of claim 4, wherein the common features used during the automatic sorting include one or more of a group consisting of size, brightness, color, shape, texture, moment of inertia, context, proximity to wafer features, proximity to other defects, connectivity to adjacent features, connectivity to other defects, and yield relevant properties derived from the corresponding image.

11. The method of claim 1, wherein the common features used during the automatic sorting include defect coordinates in wafers.

12. The method of claim 1, wherein the common features used during the automatic sorting include defect coordinates when spatial cluster analysis is used.

13. The method of claim 1, wherein common features used during the automatic sorting include information derived from one of the processing history, yield relevance, and origins of defects.

14. The method of claim 1, wherein automatically sorting the working set of images includes using a Kohonen map technique.

15. The method of claim 14, wherein the Kohonen map is seeded with non-random numbers.

16. The method of claim 14, wherein displaying the groupings includes arranging the working set of images to reflect the Kohonen map'"'"'s layout.

17. The method of claim 14, wherein displaying the groupings includes arranging the working set of images into natural groupings or clusters.

18. The method of claim 1, wherein automatically sorting the working set of images includes using a K-means technique.

19. The method of claim 1, wherein automatically sorting the working set of images includes using a spatial signature analysis technique.

20. The method of claim 1, wherein automatically sorting the working set of images is based on a plurality of cluster features that each represent a cluster of the working set of images.

21. The method of claim 4, further comprising receiving input from the user or automatically receiving input for selecting a number of features to use for the automatic sorting and/or classification and applying such selected feature number to the automatic sorting and/or classification.

22. The method of claim 1, wherein the common features used during the automatic sorting include tool history information relating to an inspection system or tool history information relating to the past success rate of the classification step.

23. The method of claim 1, wherein the working set of images originate from a semiconductor inspection process.

24. A system for classifying a plurality of images, comprising:

a software portion configured to provide a working set of images;

a software portion configured to prior to a user performing any classification of the working set of images, automatically sort the working set of images into a plurality of groupings based on common features of the working set of images and displaying such groupings; and

a software portion configured to after automatically sorting the working set of images into groupings, receive input from the user to manually classify at least a subset of the working set of images facilitated by the displayed groupings.

25. The system of claim 24, wherein the groupings are initially displayed as a plurality of elements, wherein each element specifies a number that indicates how many of the working set of images are grouped together.

26. The system of claim 25, wherein each element is selectable by the user to thereby display the corresponding one or more working set of images that are grouped together.

27. The system of claim 24, a software portion configured to:

provide a training set, wherein the training set is formed from the user'"'"'s manually classified subset of the working set of images; and

automatically classify the unclassified working set of images based on a plurality of features extracted from the training set and the working set of images and the user'"'"'s manual classification of the training set.

28. The system of claim 27, a software portion configured to display a visual representation of a comparison between the automatic classification and the manual classification performed by the user.

29. The system of claim 28, a software portion configured to receive input from the user to alter the training set based on the displayed visual representation of the comparison between the automatic classification and the manual classification performed by the user so that the automatic classification more closely matches the manual classification.

30. The system of claim 28, a software portion configured to receive input from the user to alter one or more parameters of the automatic classification based on the displayed visual representation of the comparison between the automatic classification and the manual classification performed by the user so that the automatic classification more closely matches the manual classification.

31. The system of claim 30, wherein altering the one or more parameters of the automatic classification includes graphically manipulating one or more images.

32. The system of claim 27, a software portion configured to automatically classify a second working set of images based on a plurality of features extracted from the training set and the second working set of images and the user selected classes of the training set.

33. The system of claim 27, wherein the common features used during the automatic sorting include one or more of a group consisting of size, brightness, color, shape, texture, moment of inertia, context, proximity to wafer features, proximity to other defects, connectivity to adjacent features, connectivity to other defects, and yield relevant properties derived from the corresponding image.

34. The system of claim 24, wherein the common features used during the automatic sorting include defect coordinates in wafers.

35. The system of claim 24, wherein the common features used during the automatic sorting include defect coordinates when spatial cluster analysis is used.

36. The system of claim 24, wherein common features used during the automatic sorting include information derived from one of the processing history, yield relevance, and origins of defects.

37. The system of claim 24, wherein automatically sorting the working set of images includes using a Kohonen map technique.

38. The system of claim 37, wherein the Kohonen map is seeded with non-random numbers.

39. The system of claim 37, wherein displaying the groupings includes arranging the working set of images to reflect the Kohonen map'"'"'s layout.

40. The system of claim 37, wherein displaying the groupings includes arranging the working set of images into natural groupings or clusters.

41. The system of claim 24, wherein automatically sorting the working set of images includes using a K-means technique.

42. The system of claim 24, wherein automatically sorting the working set of images includes using a spatial signature analysis technique.

43. The system of claim 24, wherein automatically sorting the working set of images is based on a plurality of cluster features that each represent a cluster of the working set of images.

44. The system of claim 27, a software portion configured to receive input from the user or automatically receiving input for selecting a number of features to use for the automatic sorting and/or classification and applying such selected feature number to the automatic sorting and/or classification.

45. The system of claim 24, wherein the common features used during the automatic sorting include tool history information relating to an inspection system or tool history information relating to the past success rate of the classification step.

46. The system of claim 24, wherein the working set of images originate from a semiconductor inspection process.