Bounding box

Updated 1 week ago by Alex Cota

Overview

Bounding boxes are the simplest form of annotation you can place on an image, created by marking the two opposite corners of a box. Bounding boxes in Labelbox are designed to be easy to create and manipulate.

Configuring bounding box

Nested classifications

Any existing classification tools can be nested within the tool to capture metadata about the label. The example below demonstrates how to nest a single-choice radio question within a bounding box object.

Creating instances

From the labeling interface, you can use the same class for more than one annotation. For example, if there are 2 birds in an image and you would like to assign the "Bird" class to all 5 birds, you can create multiple instances of the "Bird" class.

Follow these steps to create multiple instances of the same object:

  1. Select a class and draw the object.
  2. Select the same class again.
  3. Draw the next instance of the object.

Label format

Bounding box labels can be imported as an existing label or a model prediction. They can also be exported in JSON format. The coordinates in the geometric key are in the clockwise direction and the last coordinate is not the same as the first.

"Label": {
"classifications": [],
"objects": [
{
"bbox": {
"height": 272,
"left": 111,
"top": 137,
"width": 324
},
"color": "#8000FF",
"featureId": "ck2f6x8go8vcb0944ytqlq9km",
"schemaId": "ck1cnd3zpzqna070145avuefj",
"title": "Bird",
"value": "bird"
}
]
}

Was this section helpful? Give your feedback below.


How did we do?