AI Facial Recognition

Papirfly’s AI Facial Recognition feature automatically detects and tags people in image assets. This reduces manual tagging and makes it easier to find assets featuring specific individuals, like employees, photoshoot models, or event attendees.

When enabled, the system can be configured to scan both newly uploaded and existing assets, indexing faces to keep your archive organized and searchable as it grows.

Who can use it?

Facial recognition is available on request. Please contact your Customer Success Manager to have it enabled for your organisation.

Once enabled:

• Facial analysis becomes available across all archives. 
• Users with Archive Settings permissions can choose whether analysis runs automatically on all new uploads — recommended for archives that primarily host photography or people imagery. 
• Auto-run on upload is off by default to help manage API usage and associated costs.

• Users need View AI tags permission to see recognized faces.
• Users need Edit AI tags permission to add, change, or remove face names.

How does it work

Once enabled, you can configure any archive to have facial recognition auto-run on all uploads. This allows you to have control over it runs. e.g. have auto-run turned off for archives that predominantly store documents. 

This feature currently works with images, and requires a file size under 15mg. If the original is over 15mg, a derivative of less than 15mb will be used. 

The system identifies faces and suggests potential matches based on existing tagged individuals, while users can verify these suggestions and link faces to specific names.

Organizations in many sectors use face recognition in their DAM for different purposes:

• Media organizations use it to find images of specific public figures quickly and keep photo archives organized.
• Corporate communications teams use it to manage employee headshots and event photography.
• Educational institutions use it to organize student and faculty images and maintain historical photo archives.
• Marketing and content teams use it to keep content fresh. If the same model or employee appears too often across campaigns, you can see every image of that person at a glance and pick an alternative.
• Organizations managing employee departures use it to find every image of a person who has left and archive them in one go. Combined with the asset usage log, you can see everywhere those images have been used, so you can remove or replace them across your documents and web pages.

Searching 

1. Navigate to your archive
2. Open the AI detection filter 
   
   
    
3. Scroll and click to filter by the recognized face, or type in the search to find a specific name.
   1. Note: Within Place, you will be able to filter by recognized faces that have been identified but not named. This streamlines searching & naming for DAM administrators
   2. Filters across the rest of the Papirfly Suite including Point, Produce etc. will only display recognized faces that have been named.
4. Click on the recognized face you want to view
5. You'll be directed to a filtered view showing all the assets where this person's face appears.

Identifying faces 

Note: Ensure you have adequate consent before processing and using this feature. 

1. Navigate to your archive
2. Click to view the asset detail of an asset
3. Scroll to the AI tab
4. If your archive is not configured to auto-run facial recognition on upload, you can manually run facial recognition on this tab 
   
   
    
5. The recognition will identify and automatically group faces that are the same. These will show as "unnamed" faces here and in the place filter until you give the face a name 
    
6. To name the face, you must be in edit mode of the asset, once here you can click inside the text box next to a face and give it a name. You only need to name a face once, and this name will be applied everywhere that face is shown.
7. To help with identification, when you click on a face in the image, the panel on the right will automatically scroll to and highlight that person’s name. The reverse also works: clicking a name in the right-hand panel will highlight the corresponding face in the image. 
   
   
8. Click save to apply the changes 

To read more about how you can harness AI to streamline your consent management, refer to Facial Recognition & GDPR Management. 

FAQs

Are celebrities or stock images automatically recognized and named?

No, celebrities and stock image faces will be identified, but will need to be defined and named in the same way of any other face 

Can I reuse any tagging i've done manually before this feature?

There is no option to import such data in bulk 

How many faces can be identified in an image?

A maximum of 100 faces can be found in a single image 

Can you have two names on one face? 

No, you cannot 

Is there a character limit for names? 

Yes there is a limit of 255 bytes 

How do you handle the data used for Facial Recognition?

You can refer to our privacy policy for more information