2D Segmentation of Concrete Samples for Training AI Models


This web-based validation system has been designed to perform visual validation of automated multi-class segmentation of concrete samples from scanning electron microscopy (SEM) images. The goal is to segment automatically SEM images into no-damage and damage sub-classes, where the damage sub-classes consist of paste damage, aggregate damage, and air voids. While the no-damage sub-classes are not included in the goal, they provide context for assigning damage sub-classes.


The motivation behind this web validation system is to prepare a large number of pixel-level multi-class annotated microscopy images for training artifical intelligence (AI) based segmentation models (U-Net and SegNet models). While the purpose of the AI models is to predict accurately four damage labels, such as, paste damage, aggregate damage, air voids, and no-damage, our goal is to assert trust in such predictions (a) by using contextual labels and (b) by enabling visual validations of predicted damage labels.

Access to visual validation tools

Access to data downloads

  1. RAW: 1360 FOVs from 1 Fully Sampled Core (Sample M3_Day180)
  2. zip file download (674 MB)
  3. MANUAL MASK: 12 FOVs manually annotated with pixel-level annotations of 9 contextual classes
  4. zip file download (344 KB)
  5. ASSISTED MASK: 106 FOVs with damage-assisted annotations
  6. zip file download (1.2 MB)
  7. ASSISTED MASK: 1360 FOVs with context-assisted annotations
  8. zip file download (20.3 MB)
    Note: The labels assigned manually to these FOVs in order to subselect 1323 accurate FOVs can be obtained by downloading the CSV file from the visual validation tool here.


If you use this data then, please, cite the appropriate publication:

Approaches to Training Multi-class Semantic Image Segmentation of Concrete Cancer
Peter Bajcsy1, Steve Feldman2, Michael Majurski1, Kenneth Snyder2, and Mary Brady1
1Information Technology Laboratory, NIST
2Engineering Laboratory, NIST
(under review)