Photomask Defect Detective

Inspiration

Semiconductor Manufacturing Optics play a key role in structuring microchips, photomasks sometimes have production defects. These defects must be detected and visualized for the user.

What it does

Photomask Defect Detective provides a user interface to upload a tiff file which then gets transported to the python Deep Learning module where the image gets classified

How we built it

Given a set of pictures

Challenges we ran into

CORS in Azure, Max File Upload Size Azure (CS)

Accomplishments that we're proud of

DL Model

What we learned

Azure Infrastructure, Image classification, Photomasks

What's next for Photomask Defect Detective

Automatic repair module for the defected photomask

Built With

• angular.js
• azure
• azure-notebooks
• keras
• labelbox
• machine-learning-service-workspace
• primeng
• python
• redux
• sass
• tenserflow
• typescript