Facial Recognition is an AI-powered feature that detects faces and helps associate people across your workspace. This feature streamlines the process of organizing and searching for assets containing specific individuals, saving your team significant time on manual tagging and improving overall asset discoverability. Air designed Facial Recognition to keep you in control. The system suggests identifications, but workspace admins validate connections and maintain oversight.

With Facial Recognition enabled on your Air workspace, you'll see some amazing benefits.

Use Facial Recognition to get a gallery view of identified people in your workspace default sorted by first name, access the images they’re tagged in, and see what images are being used to train the AI model.

The People area shows identified people and the images used to train recognition in your workspace. Coverage improves as your team adds confirmed key images for each person.

People in this area include those created or associated manually, as well as those associated with faces Air detects.

Use the edit menu for any person thumbnail to view all the images found of them in your workspace. Manually tagging a few clear examples creates key images, and Air uses those confirmed examples to expand matches for that person across other ready images in your workspace.

Confirming a face Air has detected creates a key image—a reference used to recognize that person.

You're able to change the name of any person in your workspace and also include their title.

Train Facial Recognition to work for your specific workspace and get better as you add more content and key images. A key image is created when you confirm who a face detected by Air belongs to, helping train further matches. Manually associating a person when no face is detected links them to the asset without creating a recognition reference, whether you tag the whole image or draw a region. Key images are the confirmed examples Air uses to learn a person in that workspace. Adding the first few accurate tags helps recognition expand to additional ready images over time. Clear, accurate key images help improve matching quality, while incorrect key images can lead to broader or incorrect matches. The selected images can be reviewed and removed with ease!

Head to a view of all photos for a specific person identified in your workspace. The view automatically has a filter applied.

Get rid of a person that's been tagged in your workspace in a snap. Deleting a person won't delete their images from your workspace, just their name from wherever it has been applied.

You can label detected faces or manually associate an existing or new person even when no face is detected. A whole-image tag or manually drawn region associates the person with the asset, not a searchable recognition face. See Managing people for the workflow.

If any faces are detected in an image, you should see a magnifying glass on the top left of the asset preview:

The People panel lists detected faces and their associated names. Manual association also supports people whose faces Air has not detected.

Filter for a specific person and leverage your other filters too. Show all approved assets where that person is tagged. Show all images containing that person from a recent event you have a board for or identified in your workspace using a custom field. With Facial Recognition built into filters, finding what you're looking for just got a whole lot easier.

If Air misses a face, you can manually associate the person with the asset. If manual association controls are unavailable, contact Air support using the messenger chat on this article page or email help@air.inc. Air can review whether the issue is related to a known Facial Recognition problem, missing workspace enrichment for Facial Recognition, or plan availability.

Want to see Facial Recognition in your workspace? Learn how Air's AI Features can increase the impact of your team's creative work. Send us a note using the chat bubble below or email us at hello@air.inc to get started.