Patent · US Active

Method of generating a training set usable for examination of a semiconductor specimen and system thereof

US10832092B2 · kind B2 · utility

6Cited by
1References
11Claims
0Family size

Assignee

Inventors

Key dates

Filing dateFeb 7, 2019
Grant dateNov 10, 2020
Priority date
Expiry dateFeb 7, 2039

Classification

  • Technology area (CPC G)Physics
  • CPC primaryG06T2207/30148
  • WIPO fieldComputer technology
  • WIPO sectorElectrical engineering

Abstract

There is provided a method of examination of a semiconductor specimen. The method comprises: upon obtaining by a computer a Deep Neural Network (DNN) trained for a given examination-related application within a semiconductor fabrication process, processing together one or more fabrication process (FP) images using the obtained trained DNN, wherein the DNN is trained using a training set comprising synthetic images specific for the given application; and obtaining, by the computer, examination-related data specific for the given application, and characterizing at least one of the processed one or more FP images. Generating the training set can comprise: training an auxiliary DNN to generate a latent space, generating a synthetic image by applying the trained auxiliary DNN to a point selected in the generated latent space, and adding the generated synthetic image to the training set.

Source: USPTO / EPO open patent data. Objective bibliographic and citation counts.