Automatically labeling objects

The auto-label function uses the labels in the deployed model to generate new labels in the data set, which increases the number of labeled images in the data set. The updated data set can be used to train a new, more accurate model.

• You can automatically label images or videos where labels were not manually added. If any labels were manually added, that image or frame is skipped.
• Any automatically added labels that are saved or edited are converted to manual labels.
• If images or frames have labels that were added by the auto-label function, those images and frames are reprocessed. The previous labels are removed, and new labels are added.
• If you use a trained Detectron or high-resolution model and enabled the segmentation option to generate the labels, polygons are used instead of rectangular boxes.
• When also augmenting images, it is recommended that you accept or reject labels before you augment the data set because auto-label confidence levels are not preserved in augmented images.