Deep learning for semantic segmentation of patterns

摘要

In this specification, a method of training a deep learning model of a patterning process is described. The method comprises (i) training data including an input image and an actual image of at least a portion of a substrate having a plurality of features, and (ii) a set of classes, each class corresponding to a feature of a plurality of features of the substrate in the input image. , And (iii) obtaining a deep learning model configured to receive a set of training data and classes; Generating a predicted image by modeling and/or simulation of a deep learning model using an input image; Assigning a class of the set of classes to a feature in the predicted image based on matching of the feature and the corresponding feature in the actual image; And by modeling and/or simulation, generating a trained deep learning model by iteratively assigning weights using a loss function.