System and method for enhancing defect detection in optical characterization systems using a digital filter

A system, the system comprising:

• a controller including one or more processors configured to execute a set of program instructions stored in memory, the set of program instructions configured to cause the one or more processors to:
• acquire two or more sample images, wherein one or more difference images are generated by calculating pixel-by-pixel differences between the two or more sample images acquired, the two or more sample images including two or more photomask images, the one or more difference images including defect data, wherein the defect data includes one of defect of interest data or nuisance event data;
• generate one or more filtered images by applying a digital filter to each of the one or more difference images, the digital filter including a convolution filter including one or more convolution filter coefficients;
• and
• adjust the one or more convolution filter coefficients of the applied digital filter using a machine learning classifier to generate an adjusted digital filter, the adjusted digital filter including one or more bias coefficients and one or more adjusted convolution filter coefficients configured to enhance defect detection of an inspection sub-system by enhancing defect signals and suppressing at least one of noise signals or nuisance signals, wherein the machine learning classifier is continuously trained in real-time during operation of the system using inspection data to further update the one or more adjusted convolution filter coefficients and the one or more bias coefficients to increase separation between the defect signals and at least one of the noise signals or the nuisance signals.

A method, the method comprising:

• acquiring two or more sample images, wherein one or more difference images are generated by calculating pixel-by-pixel differences between the two or more sample images acquired, the two or more sample images including two or more photomask images, the one or more difference images including defect data;
• generating one or more filtered images by applying a digital filter to each of the one or more difference images, the digital filter including a convolution filter including one or more convolution filter coefficients;
• and
• adjusting the one or more convolution filter coefficients of the applied digital filter using a machine learning classifier to generate an adjusted digital filter, the adjusted digital filter including one or more bias coefficients and one or more adjusted convolution filter coefficients configured to enhance defect detection of an inspection system by enhancing defect signals and suppressing at least one of noise signals or nuisance signals, wherein the machine learning classifier is continuously trained in real-time during operation of the system using inspection data to further update one or more adjusted convolution filter coefficients and the one or more bias coefficients to increase separation between the defect signals and at least one of the noise signals or the nuisance signals.