I am reading an article about OPC model calibration, entitled "The new test pattern selection method for OPC model calibration, based on the process of clustering in a hybrid space".
I am curious to understand the relation between OPC calibration and the sampling method of test pattern. This article talks about a sampling method using centroid based clustering in a hybrid space. The algorithm in itself is quite mathematical, however the idea behind it i think is to choose the best test patterns sampling plan that optimizes the prediction of the OPC model calibration. The difference between the test structure on the mask and on the final pattern on the wafer should be minimal. The OPC prediction should be at its best.
Is that a good understanding for the sampling method for OPC Calibration? Am I missing something?