Optical metrology with nuisance feature mitigation

An optical metrology device configured for optical measurement of a sample, the optical metrology device comprising:

• a light source configured to generate light to be incident on a target structure on the sample, the target structure comprising a structure-of-interest (SOI) having a set of known parameters and an unknown structure that has unknown parameters;
• at least one detector configured to detect light from the target structure produced in response to the light that is incident on the target structure;
• and
• at least one processor coupled to the at least one detector, wherein the at least one processor is configured to:
• obtain metrology data from the detected light, wherein the metrology data is a combination of a first contribution from the SOI and a second contribution from the unknown structure;
• and
• determine a first set of parameter values characterizing the SOI and a second set of parameter values characterizing the unknown structure as predicted by a trained neural network using the metrology data.

An optical metrology device configured for optical measurement of a sample, the optical metrology device comprising:

• a light source configured to generate light to be incident on a target structure on the sample, the target structure comprising a structure-of-interest (SOI) having a set of known parameters and an unknown structure that has unknown parameters;
• a detector configured to detect light from the target structure produced in response to the light that is incident on the target structure;
• a means for obtaining metrology data from the detected light, wherein the metrology data is a combination of a first contribution from the SOI and a second contribution from the unknown structure;
• and
• a means for determining a first set of parameter values characterizing the SOI and a second set of parameter values characterizing the unknown structure as predicted by a trained neural network using the metrology data.