Labelbox documentation

Getting started

Hi and welcome to Labelbox!

This section contains beginner tutorials for new Labelbox users. These end-to-end exercises will show you the basic functionality of Labelbox by taking you through some simple, yet common, workflows in Labelbox.

You can find our more advanced instructional guides in our How-to guides section.

We also have a set of tutorials specific to our Python SDK. You can find those Python SDK tutorials here.

Sample labeling workflow tutorial

This beginner tutorial will take you through the basic functionalities in Labelbox.

  1. Go to Labelbox and sign in.

  2. Once you are signed in, click Create project.

  3. Under Project info, enter the following information:

    1. Name: “My First Project”

    2. Description: “Learning how to set up a project in Labelbox”

    3. Then, click Next.

  4. Under Choose data, do the following:

    1. Click Add data.

    2. Click Choose file to upload and select 1-5 sample images from your computer (PNG or JPG format).

    3. Name the dataset "My first dataset" and click Start upload.

    4. The dataset should automatically be attached to the project. Click Next.

  5. In the Configure Editor step, do the following:

    1. On Editor, click the Setup button.

    2. Click Add object.

    3. Enter “Sample object 1” as the class name and select Bounding box from the dropdown menu.

    4. Click on the color dot next to Sample object 1 and enter #FFB31C as the color.

    5. Click Add object again.

    6. Enter “Sample object 2” as the class name and select Segmentation from the dropdown menu.

    7. Click Add classification and for the instructions enter “Is it daytime?”

    8. Toggle on the Required option and leave Searchable on.

    9. Under options, enter “yes”.

    10. Click Add option, and enter “no”.

    11. Click Done.

    12. Click Confirm to go back to the project setup. Then, click Next.

  6. In the Select settings step, do the following:

    1. Toggle on the Review step option. Set the Coverage to 50% and click Confirm changes.

    2. Click Finish.

  7. To label your data in the Editor, do the following:

    1. From the Overview tab, click Start labeling.

    2. From the Tools menu, select Sample object 1 and draw a bounding box anywhere on the image.

    3. Select Sample object 2 and draw a Segmentation mask anywhere on the image. To complete the shape, click on the first point.

    4. Under Is it daytime? select yes.

    5. Click Submit.

    6. Complete steps 24-28 for the remaining assets in the dataset. Then click Go to project overview.

  8. To export your labels, do the following:

    1. Go to the Export tab and click Generate export.

    2. From the Tasks menu, download My First Project: Labels export.

    3. Open the exported JSON file in any code editor.