Generating a training set usable for examination of a semiconductor specimen

A computerized system of generating a training set for training a Deep Neural Network (DNN) usable for examination of a specimen, the system comprising a processor and memory circuitry (PMC) configured to perform operations comprising:

• obtaining a group of training images each representative of at least a portion of a specimen;
• for each given training image in the group:
• i) generating a first batch of training patches, including:
• cropping the given training image into a first plurality of original patches according to a plurality of first positions on the given training image;
• and
• augmenting at least part of the first plurality of original patches in order to simulate variations caused by a physical process of the specimen, giving rise to a first plurality of augmented patches constituting the first batch of training patches;
• and
• ii) generating a second batch of training patches, including:
• shifting the plurality of first positions to a plurality of second positions on the given training image, thereby obtaining a second plurality of original patches corresponding to the plurality of second positions, and
• repeating the augmenting to the second plurality of original patches to generate a second plurality of augmented patches, thereby giving rise to a second batch of training patches;
• and
• including at least the first batch and the second batch of training patches corresponding to each given training image in the training set.

A non-transitory computer readable storage medium tangibly embodying a program of instructions that, when executed by a computer, cause the computer to perform a method of generating a training set for training a Deep Neural Network (DNN) usable for examination of a specimen, the method comprising:

• obtaining a group of training images each representative of at least a portion of a specimen;
• for each given training image in the group:
• i) generating a first batch of training patches, including:
• cropping the given training image into a first plurality of original patches according to a plurality of positions on the given training image;
• and
• augmenting at least part of the first plurality of original patches in order to simulate variations caused by a physical process of the specimen, giving rise to a first plurality of augmented patches constituting the first batch of training patches;
• and
• ii) generating a second batch of training patches, including:
• shifting the plurality of first positions to a plurality of second positions on the given training image, thereby obtaining a second plurality of original patches, and
• repeating the augmenting to the second plurality of original patches to generate a second plurality of augmented patches thereby giving rise to a second batch of training patches;
• and
• including at least the first batch and the second batch of training patches corresponding to each given training image in the training set.