Most object detection tasks that are done today are done through the
use of bounding boxes. Marking out an object by putting a rectangle over
it is a very common, time-intensive task in supervised learning. With
this tool, we hope to make the process quicker than what it has been
before.
The tool can be selected by pressing the
...icon in the toolbar, or by pressing “R” .
You can easily start to create a bounding box by left-clicking (or pressing “A”)
anywhere in the image to place the first corner. With the first corner
placed, you can move your mouse around. When you find where you want to
place the second corner, you left-click (or press “A”) again. This creates a bounding box.
Editing a bounding box
To edit a bounding box, you first have to select it using the
move/edit tool. When one or more bounding boxes are selected, you can
edit them by left-clicking on one of the corner or side points and then drag that point to the desired location.
Delete bounding boxes
See delete objects
After labelling 10 images with bounding boxes and setting their image status to "done" or "to review", we will start to train our
Object Detection Assistant. This AI-powered tool will help you create annotations quicker.