Basics

Grasp the basics to get the most out of SuperAnnotate.

Teams

A team is a group of users that work on projects. A user can be an Annotator, QA, Customer, or Viewer. SuperAnnotate allows you to manage teams within the platform. Invite members to your team, assign them roles and tasks, and track their progress.

Projects

A project is a unit that contains all the annotations and meta information for a dataset. Manage one or several projects at a time within SuperAnnotate. Process large amounts of data without compromising the annotation speed or quality.

Classes and attributes

Classes are categories to which your annotated objects belong, e.g., person or car. Make your annotations more detailed by assigning attributes to your instances, e.g., tall (person) or sports (car).

Segmentation

SuperAnnotate features advanced segmentation tools that allow you to segment objects in images and videos.

Video segmentation works the same way as image segmentation with an additional step. When you upload a video, the platform splits it into multiple frames or images.

Manual segmentation

Use advanced tools to segment your images manually. Pixel and vector projects have different sets of tools. Whether the object you’re working with is a circle, square, or has an irregular shape, you’ll find the right tool to segment it perfectly.

Automatic segmentation

  • Smart segmentation: Run your image through the smart segmentation tool to automatically divide it into multiple segments and simplify the object annotation process. This feature works for pixel projects only.

  • Smart prediction: Run smart prediction on your data to cut down annotation time. Select an image or multiple images, click on the polygon icon, and choose your preferred segmentation model for your pixel or vector project.

Smart segmentation

Annotation

SuperAnnotate’s mission is to simplify and speed up the annotation process without compromising the annotation quality. Click on the objects or tap the spacebar to assign them a class. You can speed up the annotation process by uploading pre-annotated data.