You can upload image files to Image (Vector) Projects and Image (Legacy) Projects.

Upload methods

There are a few ways to upload images into your projects.

Upload from computer

To upload images from your computer:

1. In Data, click Add.
2. Select Upload Images.
3. Drag and drop or choose files from your computer.
4. Click Upload.

Import from integration

You can add images to your Image Projects from your integrated cloud storages (AWS, GCP, Azure and Custom). SuperAnnotate will have read-only access to your data; your items are stored in your cloud storage and are only displayed in SuperAnnotate.

Learn how to create integrations:

• AWS
• GCP
• Azure
• Custom

To add images from your integrated cloud storage to SuperAnnotate:

1. Go to Data.
2. Click Add.
3. Select Upload Images.
4. Go to the External storage tab (Image Projects and Image (Legacy) Projects must be empty or newly created for this tab to show as an option).
5. In the Integrations tab, select an integration from the drop-down list.
6. Enter a folder path (optional).
7. Click Upload.

The storage’s folder structure will be replicated in SuperAnnotate’s two-level folder structure.

📘

What is a folder path?

A folder path indicates the folder that contains the items and subfolders you want to add to SuperAnnotate. So, if you type Folder A / Folder B, it means that the items will be added from Folder B, which is a sub-folder of Folder A. If you don’t type a folder path, the bucket’s items and folders (and the items inside) will be added from the root.

🚧

External storage

Once you upload items using external storage, other direct upload options will be unavailable.

Import from S3 bucket

To import images from your S3 bucket:

1. In Data, click Add.
2. Select Upload Images.
3. Go to the Import from S3 Bucket tab.
4. Fill in the following fields: access key ID, secret access key, bucket name, and folder name. The folder name is optional.
5. Click Test to see whether you have access to the S3 bucket or not.
6. If you have access to the S3 bucket, click Start.

🚧

If the image upload limit has been reached, the exceeding images won't be uploaded.

Import with SDK

To upload images from a local folder:

sa.upload_images_from_folder_to_project(
    project = "Project Name",
    folder_path = "./data/images")

To upload images to a folder in a project:

sa.upload_images_from_folder_to_project(
    project = "Project Name/Folder Name",
    folder_path = "./data/images")

📘

The syntax of project and folder paths is used across all SDK functions.

Upload specifications

Here are all the specifications you need to know about regarding uploading Image files.

🚧

Character Limitation

• Please limit your item name to 120 characters. Uploading any items whose names exceed this limitation will result in their names cropped. This can be an issue if you have any pre-annotations or JSONs relying on the original filenames.
• An item name can't contain the following characters: / \ : * ? " < > |

Upload quantity

Here are some things to keep in mind regarding the upload quantity:

• You can upload up to 2,000 images at a time. We recommend you upload up to 1,000 images at a time for a faster and smoother performance.
• You can't upload more than 50,000 images, whether it's in a root folder or a folder. The remaining images will be skipped.
• If you want to upload more than 50,000 images to your project, you can do so by creating multiple folders. For example, if you need to upload 100,000 images, you can create two folders and upload 50,000 images to each folder.

Supported file formats

You can find all the supported file formats here.

Maximum image file size and resolution

Image skipped

An image has the Skipped status if:

• It has an existing name.
• Its resolution exceeds 100 MP for Image (Vector) Projects and 4 MP for Image (Legacy) Projects .
• Its size exceeds 100 MB.
• The image upload limit has been reached.

Image failed

An image has the Failed status if:

• Its size exceeds 100 MB for Image (Vector) Projects and Image (Legacy) Projects.
• Its file format isn't supported.