Advanced Autotag Usage

Here are some advanced Autotag features that you can use to enhance your Autotag experience.

Adding to an Autotag from the grid view

If you see items that you think are relevant to your Autotag in the Nucleus grid view, and you would like to add it to your Autotag's positive training set, you can select those items. In select drop down, click into the Autotag, and click on Add to existing image Autotag.


Add to Autotag

This will bring up a modal that lets you choose which existing Autotag to add your selection to. This is equivalent to selecting items during refinement, or selecting items when creating your Autotag, and adds your selected images to the same pool. You will need to recommit your Autotag for these new items to take effect in tagging items in your dataset.

Refinement Display Settings

There are two settings you can use to enhance your refinement displaying experience. To open these settings, click on the "Display" button on the top left. This will open the following modal:

The first setting, Maximum Images Shown in Refinement lets you tune how many items you want to return in each refinement step. The default is 16, and this value is capped at 100. You can return more items if you are planning to go through many refinement steps for your Autotag, and want to spend less time waiting for Nucleus to return items for each refinement step.

The second setting, Show Full Resolution is useful if your source images are at a high resolution. By default, Nucleus downsizes all image and object thumbnails shown in Autotag to reduce network load for downloading image data, and improve image rendering speed. If you find that the items seem too downsampled, you can turn on this setting to show full resolution items.