Label Images for Object Detection using VoTT

Learn how to use VoTT (Visual Object Tagging Tool) to label images for object detection to be used within Model Builder.

Create a new VoTT Project

  1. Download VoTT (Visual Object Tagging Tool).

  2. Open VoTT and select New Project.

    VoTT Home Screen

  3. In Project Settings, change the Display Name to the name of your choosing.

  4. Change the Security Token to Generate New Security Token.

  5. Next to Source Connection, select Add Connection.

  6. In Connection Settings, change the Display Name for the source connection to a name of your choosing, and select Local File System as the Provider. For the Folder Path, select the folder that contains the training images, and then select Save Connection.

  7. In Project Settings, change the Source Connection to the connection you just created.

  8. Change the Target Connection to the same connection as well.

  9. Select Save Project.

Add tag and label images

You should now see a window with preview images of all the training images on the left, a preview of the selected image in the middle, and a Tags column on the right. This screen is the Tags editor.

  1. Select the first (plus-shaped) icon in the Tags toolbar to add a new tag.

    VoTT New Tag Icon

  2. Name the tag and hit Enter on your keyboard.

  3. Click and drag to draw a rectangle around each item in the image you want to tag. If the cursor does not let you draw a rectangle, try selecting the Draw Rectangle tool from the toolbar on the top, or use the keyboard shortcut R.

  4. After drawing your rectangle, select the appropriate tag that you created in the previous steps to add the tag to the bounding box.

  5. Click on the preview image for the next image in the dataset and repeat this process.

  6. Continue steps 3-4 for every image.

Export your VoTT JSON

Once you have labeled all of your training images, you can export the file that will be used by Model Builder for training.

  1. Select the fourth icon in the left toolbar (the one with the diagonal arrow in a box) to go to the Export Settings.

  2. Leave the Provider as VoTT JSON.

  3. Change the Asset State to Only tagged Assets.

  4. Uncheck Include Images. If you include the images, then the training images will be copied to the export folder that is generated, which is not necessary.

  5. Select Save Export Settings.

    VoTT Export Button

This export will create a new folder called vott-json-export in your project folder and will generate a JSON file named ProjectName-export in that new folder. You will use this JSON file for training an object detection model in Model Builder.

Next steps