Export data

Export the data of Pixel Projects and Vector Projects

There are two download formats for Pixel Projects and Vector Projects: SuperAnnotate and COCO.

Export in the SuperAnnotate format

To export the data of a Pixel Project or Vector Project in the SuperAnnotate format:

  1. In Download, click Prepare for download.
  2. Choose the download format: SuperAnnotate or COCO (mandatory).
  3. Check the Include fused images box to include fused images in your download (optional).
  4. Check the Pinned items box to include pinned items in your download (optional).
  5. Choose to download items by their status. Check the boxes to do so. For example, if you check the Not Started box, only the items with the Not Started status will be downloaded (optional).
  6. Click Prepare.
  7. When the download is complete, a folder will be available. Download that folder.

SuperAnnotate download folder

The downloaded folder is the same for all image statuses:

  • Classes JSON
  • Original images (if fused images are enabled)
  • Annotations JSON
  • JSON
  • Fused images (if enabled)
  • Blue masks (Pixel Projects only)
  • Folder (if available)
    • Annotation JSON
    • Fused images (if enabled)
    • Original images (if fused images are enabled)
    • Blue masks (Pixel Projects only)

🚧

A folder will only be exported if it contains images that have the image status or image statuses indicated in the download settings. The exported folder will only contain the relevant images.

📘

You can also export a list of folders via Python SDK.

Export in the COCO format

To export the data of a Pixel Project or Vector Project in the COCO format:

  1. In Download, click Prepare for download.
  2. Choose the download format: SuperAnnotate or COCO (mandatory).
  3. Check the Include fused images box to include fused images in your download (optional).
  4. Check the Pinned items box to include pinned items in your download (optional).
  5. Choose to download the images by task type. Check the boxes to do so. Learn more.
  6. Click Prepare.
  7. When the download is complete, a folder will be available. Download that folder.

📘

  • You can check one task type at a time. Only completed images can be converted into the COCO format.

  • If the completed images are in a folder, the export will contain the annotations of the completed images only.

COCO download folder

The downloaded folder contains the following items:

  • Annotations JSON
  • Annotations JSON (for folders, if available)
  • Fused images (if enabled)
  • Original images (if fused images are enabled)
  • Blue masks (Instance segmentation only)

🚧

The rotated boxes are skipped during the COCO conversion.

Export the data of Video Projects

To export the data of a Video Project:

  1. In Download, click Prepare for download.
  2. You can choose to download items by their status. Check the boxes to do so. For example, if you check the Not Started box, only items with the Not Started status will be downloaded (optional).
  3. Select Prepare.
  4. When the download is complete, a folder will appear. Download that folder.

The downloaded folder contains the following items:

  • Classes JSON
  • Video JSON
  • Folder(s) (if applicable)

🚧

A folder will only be exported if it contains videos that have the video status(s) indicated in the download settings. The exported folder will only contain the relevant video(s).

Export the data of Document Projects

To export the data of a Document Project:

  1. In Download, click Prepare for download.
  2. You can choose to download the documents by their status: Not Started, In Progress, Completed, Quality check, Returned, and Skipped. Check the boxes to do so (optional).
  3. Select Prepare.
  4. When the download is complete, a folder will appear. Download that folder.

The downloaded folder contains the following items:

  • Classes JSON
  • Document JSON
  • Folder(s) (if applicable)

🚧

A folder will only be exported if it contains documents that have the document status(s) indicated in the download settings. The exported folder will only contain the relevant document(s).


Did this page help you?