Interpolation is a process to automatically create instance labels by estimating the location that labels should be created in videos and image sequences. Interpolation uses manual labels as reference points and needs to be run separately for each object instance.
Object tracking and interpolation both automatically apply labels.
- Interpolation applies labels in between existing labels.
- Object tracking uses a single existing label and applies labels to consecutive frames.
All labels automatically created using interpolation are assigned a confidence score (α) of 99%.
Ontology shape | Interpolation | Tracking |
---|---|---|
Bounding box | ✅ | ✅ |
Rotatable bounding box | ✅ | ✅ |
Polygon | ✅ | ❌ |
Polyline | ✅ | ✅ |
Primitive | ✅ | ✅ |
Keypoint | ✅ | ✅ |
Bitmask | ✅ | ❌ |
The video tutorial below goes through the basics of using interpolation in the Label Editor.
You can use keyboard shortcuts to run interpolation.
Click the Automated labelling button in the bottom left corner of the label editor to bring up the options for automated labeling. The Tracking and interpolation section is open by default.
The Tracking and interpolation section is open by default.
Interpolation results can be improved by manually correcting some interpolated labels, and re-running the interpolation.
Enable the Interpolation auto adjustments toggle in the Drawing settings section of the label editor settings. This ensures that only labels with a confidence score (α) less than 100% are overwritten in successive interpolations, while manual labels (α = 100%) are kept and serve as key frames.
This process can be repeated any number of times.
Interpolation is a process to automatically create instance labels by estimating the location that labels should be created in videos and image sequences. Interpolation uses manual labels as reference points and needs to be run separately for each object instance.
Object tracking and interpolation both automatically apply labels.
- Interpolation applies labels in between existing labels.
- Object tracking uses a single existing label and applies labels to consecutive frames.
All labels automatically created using interpolation are assigned a confidence score (α) of 99%.
Ontology shape | Interpolation | Tracking |
---|---|---|
Bounding box | ✅ | ✅ |
Rotatable bounding box | ✅ | ✅ |
Polygon | ✅ | ❌ |
Polyline | ✅ | ✅ |
Primitive | ✅ | ✅ |
Keypoint | ✅ | ✅ |
Bitmask | ✅ | ❌ |
The video tutorial below goes through the basics of using interpolation in the Label Editor.
You can use keyboard shortcuts to run interpolation.
Click the Automated labelling button in the bottom left corner of the label editor to bring up the options for automated labeling. The Tracking and interpolation section is open by default.
The Tracking and interpolation section is open by default.
Interpolation results can be improved by manually correcting some interpolated labels, and re-running the interpolation.
Enable the Interpolation auto adjustments toggle in the Drawing settings section of the label editor settings. This ensures that only labels with a confidence score (α) less than 100% are overwritten in successive interpolations, while manual labels (α = 100%) are kept and serve as key frames.
This process can be repeated any number of times.