Navigate to "Label Definitions" tab within a project
Type the name of the new label
Select type of new label
Finish setup for label type selected (Setup will vary on label type)
Click "Save" button
Choosing a label type depends on the model being used and the intended outcome of the model.
Rectangles: Object detection
Polygons: Semantic and instance segmentation
Point: Key point detection for singular collections
Feature points: Key point detection for a group
Text: image captioning
Classification: Image classification
Multi Classification: multi-label image classification
Rectangles, Polygons, and Point labels have the option of adding sub-labels into the label definitions.
This is useful for having classes associated with higher level instances. Example: drawing rectangles around birds for object detection, but having a class for types of bird within the rectangle (Duck, Geese, etc...).
Click "Add Sublabel"
Type a sub-label name
Choose sub-label type
Fill out label options
Click on the edit icon
Make desired changes to the label options
Click on the trash-bin icon
Click "Yes" or "Confirm" to confirm deletion