Getting Started

A short introduction that will get you labeling quickly.

Step 1: Set up a project

The project is the shell of the workflow. It contains all the important components that are needed to create ground truth data for a given computer vision machine learning task.

pageCreate a new project

Step 2: Add a label configuration

The label configuration lets you configure your very own labeling toolbox with all object classes, their geometry types and additional tags and classifications.

pageWhat is a label configuration?

Step 3: Create and add datasets

Once the project is created and the labeling task is well defined, its time to upload your image datasets and add them to the project.

pageWhat is a dataset?pageManage images

Step 4 (optional): Add collaborators

Machine Learning task usually require hundreds and thousands of images, not a task that you'll be able to do on your own. This is why you can add collaborators to your projects as labelers. They will help you label images and reach your goal faster.

Step 5: Prepare your tasks

As a project administrator move the open tasks from your BACKLOG into the WAITING state to start labeling.

pageManage Tasks

Step 6: Start Labeling

Our world class labeling tool helps you label images quickly and precisely. For Polygons we are already offering our Ai assisted labeling tool that will speed up your teams work.

pageProcess a taskpageWhat is the label mode?pageAI-assisted labeling

Step 7: Export your labels

When you are done with all your images, or if you want to already pre-train networks with the images that you have labeled so far, simply press the export button and get your JSON file with all the labeled data.

pageExport data

Last updated