Video Recap

1. Select Everything - You can also press Z for a hotkey. A spinner will occur that will not allow you to use the annotator. Depending on the quality desired as well as the number of objects in the image, it will take anywhere from a few seconds to half a minute.
2. Select Classes for Suggested Masks - Now that the proposed masks have appeared on the image, you can select class tags and then select the masks, which will then assign the masks with the selected class.
3. Select Confirm to Set Annotations - Now that you have selected all your preferred annotations, select Confirm or press the Spacebar. You have now made the annotations!

How Everything Works

Selecting the tool will automatically create predicted masks for every object detected in the image. These will appear as greyed-out masks with a dashed outline. From here, you can then select the appropriate class tag and select and assign masks you deem to be appropriate and accurate. These will then appear as solid masks in their assigned tag colors. Once satisfied, you can confirm the annotations. This is a quick way that reduces annotation time. Users will see similarities with the idea of our Model-Assisted Labelling tool. However, the underlying SAM model is capable of identifying objects in a class-agnostic manner, whereas the Model-Assisted Labelling leverages your previously trained models on Nexus to label your data with the class assigned as well, given the contextual knowledge it has gathered from prior trainings.

You will also be able to select the granularity of the auto-annotation, to trade-off between speed and accuracy. We would suggest if there are objects in the image, to use the lower granularity, as the higher granularity means that the model will have more time to annotate more objects.

Common Questions

How is Everything different from Model-Assisted Labelling?

The underlying SAM model is capable of identifying objects in a class-agnostic manner, whereas the Model-Assisted Labelling leverages your previously trained models on Nexus to label your data with the class assigned as well, given the contextual knowledge it has gathered from prior trainings.