Adobe MAX Sneaks 2022: Every AI Experiment, and its Developer

Published

Kevin Hart and Bria Alexander on stage with 'Sneaks Multiverse' text projected in the background

Here's a quick and bit-size visual summary of every Adobe AI experiment from the Sneaks section of Adobe MAX 2022 (presented by Bria Alexander and the hilarious Kevin Hart). If you're just here for Kevin Hart, scroll to the end and you won't be disappointed.

Project Vector Edge

Use AI to automatically project 2D designs onto 3D surfaces 🤯. The AI can even wrap vectors around surface edges:

Just by dragging the vector, it will adapt on any surface of a 3D image:

-

Screenshot 2022-10-20 at 12.36.09.png
Project Vector Edge, presented by Ankit Phogat

Project Instant Add

Edit videos just like you'd edit images, using AI. Just annotate the position in the video with where you'd like to add your assets, and the AI will place them, cutting out hours of manual masking work:

Man on rooftop with drawings sketched around him
Place text without manually masking, and add motion graphics to video really easily.

You can even map large design assets and environments to the space on your video:

-

Screenshot 2022-10-20 at 12.21.29.png
Project Instant Add presenter, Joon-Young Lee

Project Magnetic Type

Fuse shapes to digital text in a click. Once fused with text, the shape stays magnetised to the letter it's been attached too, even as you type:

-

Screenshot 2022-10-20 at 12.30.51.png
Project Magnetic Type was presented by Arushi Jain

Project Clever Composites

Project Clever Composites uses AI to carry out all the manual tasks involved in compositing images. For example, it autoscales, analyzes background geometry and adds shadows, creating realistic composites in seconds: 

Frame 1.png
Clever Composite can place objects onto backgrounds.
Screenshot 2022-10-20 at 13.38.06.png
Project Clever Composites was presented by Zhifei Zhang

Project Motion Mix

Use AI to add body motion to stationary pictures of people:

-

You can place people in different environments and give them synchronised movements based on predefined sample dances:

-

Screenshot 2022-10-20 at 12.40.47.png
Project Motion Mix, Jae shin Yoon

Project Blink

This project uses AI to make videos searchable. It creates a transcript, tagged with audio and speakers, so you can find out who said what with just a text search!

For example, if you remember a backpack appearing in a video, you can just search the word 'backpack' and find the exact frame, as well as a transcript of what has been said at that moment:

Screenshot 2022-10-20 at 12.49.35.png
Search a video for the word 'backpack'
Screenshot 2022-10-20 at 12.47.17.png
Project Blink was presented by Mira Dontcheva

Project Artistic Scenes

Turn 2D images into 3D scenes with more crazy AI. For example, this 3D man and environment is generated from a few regular photos. The AI even created the side of the man that was not shown in the photo:

Furthermore, different textures can be applied to the 3D scene, so you can see it in different 3D styles:

manchair.png
Add different textures to 3D scene (that was originally a 2D image)
Screenshot 2022-10-20 at 12.54.45.png
Project Artisit Scenes, Sai Bi

Project All of Me

'Uncrop' your images. The AI in 'Project All of Me' can build the missing environment around a photo, thereby 'uncropping' the image:

Uncropping an image. 3 images side by side, showing how the AI analyzes a cropped image and fills in what's missing
Here the image on the left is 'Uncropped' - the rest of the image is generated by AI on the right.

It can also extend and edit images, such as increasing the length of the dress in the original. You just draw on the part you want to extend, and AI will do the rest:

Screenshot 2022-10-20 at 13.01.17.png
The dress is extended by AI

From there, different textures can be added to the dress, and it's possible to generate limitless variations:

Screenshot 2022-10-20 at 13.01.47.png
Add texture to a dress
Screenshot 2022-10-20 at 13.01.59.png
Generate variations for different dress styles and textures
Screenshot 2022-10-20 at 12.58.38.png
Project All of Me, Qing Liu

Project Made in the Shade

Uses AI to accurately add shadows to images, making it possible to move 3D objects around realistically. Existing shadows on a photo can even overlay new objects added to a scene:

You can even move the shadow - "we control the sun":

-

Screenshot 2022-10-20 at 13.07.43.png
Project Made In The Shade, Vojtech Krs

Project Beyond the Seen

Create 360 scenes from a flat photograph or image. For example, from this flat image:

Screenshot 2022-10-20 at 13.04.32.png
Original 2D photo

This 3D Scene was built, with a surrounding panoramic room generated by AI:

-

Screenshot 2022-10-20 at 13.03.14.png
Project Beyond the Seen, Yannick Hold-Geoffroy


Time to Sneak Off 🖖

That's all of the highlights, I must sneak off now, and leave you with Kevin Hart. For more information on each project, check out:

Since the developers seemed to ignore Kevin, here's some Kevin Hart MAX appreciation:

Screenshot 2022-10-20 at 12.10.32.png
Kevin Hart entering the scene
Screenshot 2022-10-20 at 12.12.33.png
Presenters Kevin Hart an Bria Alexander with a little flask
Screenshot 2022-10-20 at 12.11.35.png
Bria Alexander and K Hart
Screenshot 2022-10-20 at 12.12.22.png
Drink
Screenshot 2022-10-20 at 12.12.52.png
Multiverse 🤯
Share to your friends
Author avatar

Rob Díaz