CRS
DI
Live Experimental Analysis

Model card

Technical transparency for the crack-like region detection model.

Model
Crack-like region analysis
Status
Experimental AI POC
Human review
Required

Purpose

Segment crack-like regions in still images of concrete surfaces to accelerate visual inspection triage. Not intended for structural safety conclusions or physical measurement.

Data documentation status

Dataset composition, annotation methods, provenance, rights, and evaluation splits are in development. CRS does not publish performance conclusions until the supporting documentation and evaluation are reproducible and approved.

Evaluation status

No accuracy, IoU, precision, recall, or F1 result is approved for publication. Evaluation design, held-out testing, error analysis, and reproducibility controls remain in development.

Input boundaries

Input-quality boundaries remain under evaluation. Lighting, angle, resolution, surface finish, moisture, texture, prior repairs, and image processing can affect experimental suggestions.

Known limitations

  • No physical width measurement (uncalibrated pixel output only).
  • Confusion with surface staining, joints, and wire marks.
  • Fine, low-contrast, obstructed, or poorly lit regions may be missed or misrepresented.
  • No structural-safety or definitive crack-cause conclusion.

Known failure directions

Surface staining may be suggested as crack-like.
Joints, formwork lines, and wire marks may be confused with crack-like regions.

Version

Internal experimental concept. Public versioning documentation is in development.
AI-assisted inspection proof of concept. Outputs require professional human review and do not constitute structural engineering conclusions.