| CPC H01L 22/12 (2013.01) [G06F 18/214 (2023.01); G06N 20/00 (2019.01)] | 20 Claims |

|
15. A computerized method, the method comprising:
generating a training set comprising a plurality of training images of a semiconductor specimen and respective ground truth measurement data associated therewith pertaining to a metrology application; and
training an end-to-end (E2E) learning model for the metrology application using the training set and one or more cost functions, the one or more cost functions specifically configured to evaluate, for the plurality of training images and corresponding training measurement data outputted by the E2E learning model, one or more metrology benchmarks from a group comprising: precision indicative of repeatability of first training measurement data of first different training images of the plurality of training images acquired for a given feature on the specimen by one metrology tool, correlation between the corresponding training measurement data of the plurality of training images and the respective ground truth measurement data associated therewith, and matching indicative of repeatability of second training measurement data of second different training images of the plurality of training images acquired for the given feature by different metrology tools;
wherein the E2E learning model, upon being trained, is usable for processing a runtime image of the semiconductor specimen and obtaining runtime measurement data specific for the metrology application.
|