Adobe has been actively developing generative AI features in recent years, and at their recent Adobe MAX event, they showcased a series of experimental generative AI capabilities during the "Sneaks" segment, offering the audience a glimpse into the AI-powered creative experiences of the future.
This year's Sneaks segment at Adobe MAX, similar to last year, focused on tools that aid in creation and editing, providing users with faster, more accurate experiences and reducing unnecessary processing time.
- Project Perfect Blend: Facilitates the seamless integration of a person or object into another image by automatically adjusting color, lighting, and even shadows for a natural blend. Leveraging generative AI, it assists in creating custom backgrounds and enhances foreground masking functionality, particularly excelling in natural blending and realistic shadow casting, making the process smoother and more lifelike.
- Project Clean Machine: Enables users to easily remove unwanted flashes from photos or videos, such as camera flashes or fireworks, as well as temporarily obstructive objects. It can even detect and automatically remove flashes from videos, offering great convenience when cleaning up unnecessary elements in footage.
- Project In Motion: Allows custom-shaped animations to be converted into videos through simple content descriptions. Using generative AI, it converts text into animations based on initial After Effects projects. Users can also add reference images to blend styles with prompts, creating unique effects.
- Project Know How: With digital fingerprinting and watermark technology, this project can quickly trace the origins of videos and images, identifying creators and AI-generated records, thus enhancing content transparency.
- Project Turntable: Enables 2D vector data to rotate within 3D space while maintaining its original 2D style from any new angle. Even after rotation, vector graphics retain their original form, facilitating consistency in the original design.
- Project Super Sonic: Uses prompts to generate sound effects for videos or lets users click on objects within the video to create sound effects. Users can control the timing of these sounds by voice, or directly overlay sounds along the timeline, blending background and foreground sounds, and select from prompt options to achieve suitable sound effects.
- Project Scenic: Enables users to establish 3D scene layouts via prompts, making it easier to create 2D images. This tool allows for camera control and single-object adjustment to guide the image generation process, thereby reducing the trial and error involved in layout and camera adjustments.
- Project Remix A Lot: Makes it easy to turn a rough sketch into a refined design. Users can start with a basic idea, then use an inspiration sketch to enhance it. The tool uses generative AI to guide the creative process, allowing designs to adapt to various shapes and sizes, including irregular ones, while maintaining consistent styling across different works.
- Project Hi-Fi: Lets users capture any part of an image as a guide to quickly create high-quality AI-generated images. These images can then be effortlessly imported into Adobe Photoshop for further editing.
These experimental features are currently under development at Adobe, and while they may eventually be integrated or pivot in new directions, they illustrate how the widespread adoption of AI will transform various creative workflows and methods. Creators will need to evolve alongside these advancements.