Friday, October 31st, 2025 Posted by Jim Thacker

See Adobe’s Project Surface Swap and Project Trace Erase


Adobe has previewed a new batch of experimental graphics technologies during the popular Sneaks session at Adobe MAX 2025, its user conference.

There were 10 in total, spanning the fields of image editing and relighting, radiance fields, vector animation, video editing, sound design, and audio post-production.

However, two particularly caught our eye, partly because of what they do, but also because they both seem likely to make their way into commercially available Adobe tools.

Project Surface Swap lets you select materials in 2D images as if they were 3D scenes, while Project Trace Erase removes objects from images, along with shadows, reflections and lighting.


Project Surface Swap edits materials in photos as if they were 3D scenes
Project Surface Swap uses AI-based texture recognition to identify materials within 2D images, making it possible to select and edit them in the same way that you might in a 3D scene.

Clicking on the image selects all of the areas that contain the same material, making it possible to make edits ranging from creating simple color variants to complete material replacements.

For material replacements, users can apply a source texture to the selected regions of the image, with Project Surface Swap adjusting perspective and lighting automatically.

The video above shows examples ranging from selecting the paintwork of a car to isolating vines from the wall they are climbing up, automatically generating a pretty detailed matte.

However, perhaps the most impressive is a photo of a kitchen, from which Project Surface Swap isolates a wooden work surface, ignoring the similar-looking wooden cutting boards on top of it.

Photoshop already has a number of AI-powered intelligent selection tools, like Select Subject and Select People, so selecting materials seems like a natural extension of those capabilities.


Project Trace Erase is a true magic eraser for photos
On a similar note, Photoshop already has some pretty well-established AI-trained object removal tools, but Project Trace Erase takes the workflow to a new level.

As with the existing tools, a user paints roughly over the object they want to be removed from an image, then Project Trace Erase automatically generates a clean replacement.

The difference is that it doesn’t just replace the object itself, but removes all trace of it from the image – even in parts of the image that weren’t originally selected.

Most obviously, that means the shadows the object casts on its surroundings, its reflections in nearby objects, and even the light it casts into the scene.

Less obviously, that means its other traces: the video above shows Project Trace Erase removing a brazier from a photo, along with the smoke it is emitting.

It’s worth watching the video in full: the demos keep getting more complex, with Project Trace Erase completing in seconds what would otherwise be minutes of painstaking manual clean-up.

Again, it’s an obvious extension of the existing tools in an existing Adobe application, and one we’d expect to see in a public release in future.

So when will I be able to use this technology in Adobe software?
While Project Surface Swap and Project Trace Erase are research projects, not commercial tools, tech shown in Sneaks sessions can make its way into Adobe products quite quickly.

One of the highlights of last year’s session, AI compositing technology Project Perfect Blend, just made its way into Photoshop 27.0, in the shape of the new Harmonize feature.

Watch the full recording of the Sneaks session from Adobe MAX 2025

Read Adobe’s overview of the other technologies previewed during Sneaks session


Have your say on this story by following CG Channel on Facebook, Instagram and X (formerly Twitter). As well as being able to comment on stories, followers of our social media accounts can see videos we don’t post on the site itself, including making-ofs for the latest VFX movies, animations, games cinematics and motion graphics projects.