Back to results
US 2026/0243746 A1Published application
Automatic Defect Classification Analysis
AssigneeKLA Corporation
InventorTingzhu Liu
Life of this patent
Estimating the term…
Abstract
Automatic defect classifications and manual defect classifications of a workpiece (such as a semiconductor wafer) are assembled along perpendicular axes of a matrix. Matches between corresponding entries of the automatic defect classifications and the manual defect classifications are determined. A purity rating, a contribution rating, and an accuracy rating are determined based on the matrix.
Claims16
A method comprising:
- receiving, at a processor, a plurality of automatic defect classifications for a workpiece;
- receiving, at the processor, a plurality of manual defect classifications for the workpiece;
- assembling, using the processor, the plurality of automatic defect classifications and the plurality of manual defect classifications along perpendicular axes of a matrix;
- determining, using the processor, matches between corresponding entries of the plurality of automatic defect classifications and the plurality of manual defect classifications;
- determining, using the processor, a purity rating based on the matrix, wherein the purity rating evaluates correctness of classification for defect types;
- determining, using the processor, a contribution rating based on the matrix, wherein the contribution rating evaluates a percentage of defects that were classified;
- and
- determining, using the processor, an accuracy rating based on the matrix, wherein the accuracy rating evaluates a percentage of defects that were classified correctly.
5 more elements
A system comprising:
- a semiconductor inspection system that includes:
- a particle beam source that generates a particle beam;
- a stage that supports a workpiece in a path of the particle beam;
- and
- a detector that receives the particle beam returned from the workpiece;
- and
- a processor in electronic communication with the detector, wherein the processor is configured to:
- receive a plurality of automatic defect classifications for the workpiece;
- receive a plurality of manual defect classifications for the workpiece;
- assemble the plurality of automatic defect classifications and the plurality of manual defect classifications along perpendicular axes of a matrix;
- determine matches between corresponding entries of the plurality of automatic defect classifications and the plurality of manual defect classifications;
- determine a purity rating based on the matrix, wherein the purity rating evaluates correctness of classification for defect types;
- determine a contribution rating based on the matrix, wherein the contribution rating evaluates a percentage of defects that were classified;
- and
- determine an accuracy rating based on the matrix, wherein the accuracy rating evaluates a percentage of defects that were classified correctly.
12 more elements
A non-transitory computer-readable storage medium, comprising one or more programs for executing the following steps on one or more computing devices:
- receiving a plurality of automatic defect classifications for a workpiece;
- receiving a plurality of manual defect classifications for the workpiece;
- assembling the plurality of automatic defect classifications and the plurality of manual defect classifications along perpendicular axes of a matrix;
- determining matches between corresponding entries of the plurality of automatic defect classifications and the plurality of manual defect classifications;
- determining a purity rating based on the matrix, wherein the purity rating evaluates correctness of classification for defect types;
- determining a contribution rating based on the matrix, wherein the contribution rating evaluates a percentage of defects that were classified;
- and
- determining an accuracy rating based on the matrix, wherein the accuracy rating evaluates a percentage of defects that were classified correctly.
5 more elements
Dependencies are inferred from the claim text.