Neural network training device, system and method
AssigneeSTMicroelectronics (Rousset) SAS
InventorLaurent Bidault
Life of this patent
Estimating the term…
Abstract
A device includes image generation circuitry and convolutional-neural-network circuitry. The image generation circuitry, in operation, generates a digital image representation of a wafer defect map (WDM). The convolutional-neural-network circuitry, in operation, generates a defect classification associated with the WDM based on: the digital image representation of the WDM and a data-driven model associating WDM images with classes of a defined set of classes of wafer defects and generated using a training data set augmented based on defect pattern orientation types associated with training images.
Claims33
A device, comprising:
- image generation circuitry, which, in operation, generates a digital image representation of a wafer defect map (WDM);
- and
- convolutional-neural-network (CNN) circuitry, which, in operation, generates a defect classification associated with the WDM based on:
- the digital image representation of the WDM;
- and
- a data-driven model associating WDM images with classes of a defined set of classes of wafer defects and generated using a data set of images, wherein the data set of images includes training images and augmented images generated based on defect pattern orientation types associated with the training images.
A system, comprising:
- one or more memories;
- and
- wafer-defect-map (WDM) classification circuitry coupled to the one or more memories, which, in operation, generates a defect classification associated with the WDM based on:
- a digital image representation of the WDM;
- and
- a data-driven model associating WDM images with classes of a defined set of classes of wafer defects and generated using a training data set of images, wherein the training data set of images includes training images and augmented training images generated based on defect pattern orientation types associated with the training images.
A method, comprising:
- generating a digital image representation of a wafer defect map (WDM);
- and
- generating a defect classification associated with the WDM based on:
- the digital image representation of the WDM;
- and
- a data-driven model associating WDM images with classes of a defined set of classes of wafer defects and generated using a training data set augmented based on defect pattern orientation types associated with training images.
A non-transitory computer-readable medium having contents which configure a wafer defect map (WDM) classification system to perform a method, the method comprising:
- generating a digital image representation of a wafer defect map (WDM);
- and
- generating a defect classification associated with the WDM based on:
- the digital image representation of the WDM;
- and
- a data-driven model associating WDM images with classes of a defined set of classes of wafer defects and generated using a training data set augmented based on defect pattern orientation types associated with training images.
Dependencies are inferred from the claim text.