OMB Individually Reported

Pre-trained microscopy image neural network encoders

Low riskExact public inventory row

Description

Convolutional Neural Network encoders were trained on over 100,000 microscopy images of materials.

Detailed example

Automatically segment microscopy features given limited annotated microscopy images.

AI / analytics pattern

Computer Vision: AI that processes and interprets visual data (e.g., images and videos).

Automation level / stage

c) Deployed – The use case is being actively authorized or utilized to support the functions or mission of an agency.

Expected benefit

When deployed in downstream microscopy tasks through transfer learning, encoders pre-trained on MicroNet outperform ImageNet encoders. These pre-trained MicroNet encoders have been successfully deployed for semantic segmentation, instance segmentation, and regression tasks.

Controls / human review

ATO: Yes; PIA: Not published

Data needed

Annoted microscopy images of various materials (metals, composites, EBC/CMC, etc.)