Textile, polymer & compositeWorked example

Textile pilling grade

Predict an ISO pilling grade from a fabric image, with density and the features behind the call.

Modalities:Flatbed scanMacro photo
3.5
ISO grade
18/cm²
Pill density
0.91
Confidence

Example outputs shown for illustration. Numbers depend on your samples and protocol.

The measurement, today

Pilling assessment compares an abraded swatch to reference photographs by eye. Grades drift between assessors and labs, and the rationale is never recorded.

From image to reviewed result

  1. 1

    Upload the swatch

    Drop a flatbed scan or macro photo of the abraded fabric. No fixed rig required.

  2. 2

    Detect pills

    A segmentation pass finds and outlines individual pills, separating them from weave texture and lint.

  3. 3

    Grade to ISO

    An EfficientNet-B0 classifier predicts the modified-Martindale grade, with pilling density per cm² alongside.

  4. 4

    Explain the call

    Grad-CAM highlights the regions that drove the grade; SHAP gives per-feature and per-stage contributions.

  5. 5

    Review & log

    Accept or adjust before the result is logged with the image, grade, density, and attribution attached.

What you export

ISO grade
Pilling density
SHAP attribution
Grad-CAM overlay
Stage contributions

Scope: Assists pilling assessment and QC review. It does not replace a certified human final grade unless validated for your fabrics.

Send a sample image and a measurement goal

We will show the closest ConductVision workflow and flag what needs custom validation for your images.