Comment on page
Bounding Box (Rectangle)
Use Bounding Box labels for Object Detection
The Rectangle schema element creates an array of objects with data consisting of minimum and maximum XY coordinates. These minimum and maximum values define the boundaries of the rectangle. The XY coordinate values are relative (0.0 to 1.0) to width and height respectively.
Rectangles enable object detection.
Architectures such as R-CNN, YOLO and SSD support rectangles (also known as bounding boxes) as input. Some use cases include detecting people in photos, buildings in satellite imagery, and gunshots in audio spectrograms.
Three Cheetahs detected
- 1.Navigate to the "Label Definitions" tab of your dataset
- 2.Type label name in the "Name" field
- 3.Bounding Box is default in "Type" dropdown
- 4.Select label color from the color chooser menu
- 5.Click "Save" button
- 6.Repeat as needed for multiple labels
After you've defined a Bounding Box, you can label a rectangle:
Bounding Box also supports Sublabels:
The Bounding Box (Rectangle) schema element creates an array of objects with data consisting of minimum and maximum XY coordinates. These minimum and maximum values define the boundaries of the rectangle. The XY coordinate values are relative (0.0 to 1.0) to width and height respectively.
{
"type": "rectangle",
"data": {
"min": [0.03108985876522894, 0],
"max": [0.24836992510709158, 0.26650752787428983]
},
"children": {}
}
Once you've defined a bounding box, you can start annotating your data.
You can train a bounding box object detection model by select Bounding Box as a "Model Output Option" in your SmartML configuration.
Last modified 1yr ago