wenglor sensoric - AI Lab - How does image labeling work in the AI Lab?

In this tutorial, you'll learn how to label images in AI Lab and prepare your dataset for AI model training.

Now that your images have been uploaded, the first step is to create your classification classes. On the dataset landing page, add the classes you want to use - for example, OK and NOK. You can customize the color assigned to each class at any time to make your dataset easier to manage.

There are several ways to label images in AI Lab. The simplest method is to open an image in the full-screen view and assign a class using the keyboard shortcuts. For example, press 1 for OK and 2 for NOK to quickly annotate your images.

To speed up the process, you can also select multiple images and assign the same label to all of them at once.

For even more efficient annotation, use the filter options to narrow down your dataset. For example, filter images based on the tags imported from uniVision 3. Select all images with the OK tag and assign the corresponding Ground Truth label in a single step.

If you accidentally assign the wrong Ground Truth, don't worry - you can easily remove or change the label at any time.

Next, filter for all unannotated images. These are the images that haven't been assigned a class yet. Select them all and assign the NOK class to complete the labeling process.

That's it! Your dataset is now fully labeled and ready for training an AI model in AI Lab.