GuidesDatasets and Models

Segmentation Inputs

Prepare YOLO Segmentation or COCO Instance Segmentation inputs.

Knotest supports YOLO Segmentation polygon annotations and COCO Instance Segmentation annotations. Segmentation is instance-based: the class, instance identity, and geometry must remain aligned with the Model output.

Preparation checklist

  • Confirm polygons or masks are valid for their source image dimensions.
  • Remove degenerate geometry and annotations with missing images.
  • Preserve class IDs and Dataset split selection.
  • Confirm the Model task is Segmentation, not Object Detection.
  • Record the Model input and output interpretation exactly.

Review sample previews after import. A successful file copy does not prove that the annotation geometry or Model contract is scientifically usable.

Imported Helmet segmentation samples with polygon annotations overlaid.

Watch: Inspect segmentation inputsToggle polygon annotations and review dataset settings.Silent demo · Subtitles follow page language · Full screen for detail

Object Detection Inputs

Prepare YOLO or COCO bounding-box data and a compatible detection Model.

Oriented Bounding Box Inputs

Prepare YOLO four-corner oriented bounding-box annotations and Model output.

On this page

Preparation checklist