Adobe says Max experiments turn photos into 3D scenes and let users interview AI characters

Key points

• Project Sight Span reframes photos, paintings and illustrations as explorable 3D environments, restoring resolution once an angle is chosen
• Project Screen Test lets users talk to an AI character to shape its personality before generating animated screen tests
• Project Rinse Repeat copies Photoshop edits and reapplies them to other people or objects in the same image

Adobe has unveiled a set of experimental tools at its Max conference, including one that turns flat images into 3D scenes you can move around in. The projects, known as sneaks, are previews of features that may later reach Creative Cloud apps or the Firefly platform, and Adobe gives no guarantee that they will ship.

Project Sight Span takes any flat image, whether a photo, painting or illustration, and builds a 3D environment around it, generating objects and backgrounds beyond the original frame. The 3D view looks pixelated until you pick an angle, at which point the resolution is restored, and artistic styles and details such as tattoos stay recognisable.

Characters and sound

Project Screen Test is aimed at AI-generated characters for video. Users describe base traits and a voice, then talk to the character to flesh out its backstory, fears and movement quirks before generating animated screen tests. Project Sound Bender mixes sound effects by describing them in a prompt and adjusting them on a square canvas, while Project Rinse Repeat copies Photoshop edits and applies them to other people or objects in a photo.

Adobe research scientist Gautham Mysore told The Verge that making these interactions more intuitive also gives experienced creatives more precise control. The Verge notes that the experiments, built on Adobe's generative AI work, could eventually displace the tools and stock libraries creative professionals rely on.

Read the full story on The Verge AI →