Can you detect duplicate images, or image spam?

The Near-Duplicate Detection model is used to identify images that are so called duplicates or near-duplicates. It can be used across the following use-cases:

• Image Blacklists and Disallow lists to prevent some specific images from (re)appearing on your site or app.
• Spam and Theft prevention to prevent users from submitting the same image multiple times and other users' image.

Here are some examples of transformations and modifications typically used to try to evade duplicate detection and that are detected by the Near-Duplicae Detection model:

• Resolution, size and format changes
• Overlays, watermarks, text, logos
• Cropping and reframing
• Color changes and filters
• Rotations
• Stretched and horizontal flips
• Editing, where specific parts of the image are modified

Check out the documentation to learn more.