Home / Label types / Image

Bounding box annotation

Tight rectangles around every object your model needs to find, each with the right class label and attributes.

Street image 0087QA passed
car 0.98person 0.97traffic_light 0.99tree 0.91
carpersontraffic_lighttree
AnnotatorLabels drawn
ReviewerChecked
QAApproved

What we label

  • Boxes drawn edge to edge, no padding
  • Class labels from your taxonomy
  • Occluded and truncated flags
  • Attributes such as colour, state or brand
  • Small and crowded objects included

Where it is used

  • Retail shelf and product detection
  • Street scenes for mobility
  • Kitchen and household objects
  • Document fields and logos

How we check quality

  • Box edges checked against your pixel tolerance
  • Reviewer recounts objects to catch misses
  • Class checks on every confusing pair
  • Sample of each batch re-labelled by QA

Delivery formats

We deliver in the format your training pipeline already uses. Need something else? Send us a sample file and we will match it.

COCO JSONYOLO TXTPascal VOC XMLCSV

See bounding boxes on your own data

Send a small sample and your guidelines. The pilot is free.

Request a free pilot