Video of Rotate Object by Rob de Winter
A new Photoshop feature, built with major contributions from Adobe Research, allows users to turn a 2D layer into a 3D object, rotate the angle and perspective of the object, and then place it seamlessly back into a scene.
The new capability, Rotate Object, lets users tap into the power and precision of 3D rotation, without the complexity of creating a 3D model. It’s ideal for re-arranging the elements of an existing scene or tweaking a stock or generated image to fit into a new scene. For example, imagine lifting a car from a still image, swiveling it to face a new direction, and setting it back down on the road at just the right angle—all with lighting and shadows that fit the scene.
Rotate Object is now available in public Beta.
The Rotate Object experience
For users, adjusting 2D objects with Rotate Object is a simple process. Users can open the feature through the edit menu in Photoshop, or they can enter Transform on the Contextual Task Bar. The tool creates a low-resolution (Gaussian splat) 3D version of a 2D layer, allowing the object to be manipulated quickly in three dimensions. From there, users can rotate objects using the on-canvas controls and the Contextual Taskbar, allowing them to right-click “free rotation” to move and rotate the object. When a user clicks done, the feature upscales the image to match its original 2D resolution. From there, users have the option to harmonize their image, adjusting the lighting and shadows of the edited object to match the background.
The research and collaboration behind Rotate Object
Rotate Object is the product of years of research and deep collaboration.
Adobe Research pioneered the research ideas behind this tech and then adapted them in collaboration with Adobe Photoshop, 3D&I, and Firefly to bring it to user needs. Researchers worked on the research breakthroughs, designing and training the generative and reconstruction models under the hood that produce the 3D asset and the final upscaled image, and building the Photoshop-native renderer that enables real-time control over the 3D asset.
Teams also gathered user feedback, crafted the user experience, and integrated the feature into Photoshop.
The technology also grows out of significant advances from Adobe Research in 3D-inspired tools such as the Text to 3D feature, which enables users to generate a 3D shape from text descriptions in Stager, and Turntable, which allows users to rotate vector images inside Illustrator.
Rotate Object is an example of the newest generation of Adobe’s AI tools, which are designed to give creators more capabilities and more precise control over their work instead of replacing the creative process.
Adobe Research, Photoshop, 3D&I, and Firefly worked together to bring Rotate Object to life. Key contributors include:
Adobe Research
Core Contributors: Yiwei Hu, Zhiqin Chen, Mathieu Gaillard, Yifan Wang, Zifan Shi, Kalyan Sunkavalli
Other Contributors: Kevin Blackburn-Matzen, Nathan Carr, Iliyian Georgiev, Vova Kim, Vojtech Krs, Élie Michel, Xin Sun, Hao Tan, Dipti Thakkar, Zhixin Shu
3D&I Tech Transfer Team
J Eisenmann, Raymond Fei, Tyler Georgeson, Dominique Piche-Meunier, Nikolai Svakhin, Amanda Vu-Morris, Skylar Wurster
Photoshop
Anna Chen, David Dobish, Amy Durocher, Matt Fuerch, Mihaela Hahlov, Neerja Pancholi, Claudia Rodriguez, Tushar Turkar, Dongmei Li, Sagar Pathak, Lizzy Stellato
Design
Davis Brown, Michael Cragg, CJ Gammon, Daniel Plemmons, Daniel Robbins, Evan Shimizu
Video of Rotate Object by Rob de Winter
Wondering what else is happening inside Adobe Research? Check out our latest news here.
New Photoshop feature: Rotate 2D objects in three dimensions
TAGS
AI & Machine Learning
Machine learning and AI are essential tools in Adobe Research’s work in content intelligence, customer intelligence, document intelligence, and more.
View Projects
AR, VR & 360 Photography
Adobe Research is working to empower creators to author new immersive experiences in virtual reality, augmented reality, and 360 photography. Our work includes novel interactive editing and playback methods.
View Projects
Audio
Developing new audio processing and analysis algorithms for a variety of application areas such as pro audio, audio for video, technology for novice audio users, audio interfaces, and music information retrieval.
View Projects
Computer Vision, Imaging & Video
Automating understanding of image content, style, metadata, and intent for search, recognition, and manipulation.
View Projects
Data Intelligence
Transforming customer analytical needs into advanced analytical algorithms and services that generate high-value “advice” more autonomously.
View Projects
Graphics (2D & 3D)
Creating new algorithms and interfaces to extract, analyze, model and render 3D geometry for a variety of applications.
View Projects
Human Computer Interaction
Investigating new creativity support tools, multimodal interaction, interaction across devices and contexts, visualization, and crowdsourcing.
View Projects
