Adobe has long extended its reach into various fields, including design, photography, and multimedia. With the ongoing AI revolution, Adobe proposes exciting new features to keep their software innovative in front of challengers. In this blog post, we'll dive into some of the latest developments specifically tailored for professionals that are currently in Adobe's R&D pipeline.

Firefly, the way it is presented for testing today, only showcases four features: the generative text-to-image AI, the generative fill, text effects, and generative recolor. While they are spectacular in their own ways, the lineup from Adobe is the stuff of fiction. The company targets some of the frequent needs of visual creators and some of the most time-consuming tasks for professionals.

Upscaling

Raising the resolution of an image is a next-to-impossible task. You can have limited success by giving it some sharpness or blurring some edges. Otherwise, a low-quality image will remain a low-quality image. Firefly showcased a feature that would be capable of raising the resolution of images, resulting in a sharper and clearer picture. FireflyAI will extract missing pixels from a low-quality image lacking definition or detail.

Generative Custom Brushes

Brushes are in-game currency when it comes to software like Photoshop. Artists will spend time creating their own brushes, or money purchasing brush sets. Good quality digital brushes are just as important and precious as you would expect them to be in the carbon world.

The user would essentially create personalized brushes in two ways: based on prompts or using an image as a reference. The second version is particularly interesting. Firefly would analyze textures to create brushes that match fur, hair, neon, smoke, or whatever best fits their project.

Generative Vector Images

No AI today allows you to separate your image into layers, edit individual lines, or scale up your design indefinitely. Firefly would introduce text-to-vector generative AI, which would create editable models.

This is invaluable for illustrators and logo creators, who would have a basis to work on.

Images Based on Sketches

Ever had a client draw a design on the back of a napkin? Adobe's R&D is currently coming up with a feature that can turn those hand-drawn sketches into professional-looking photographs. The feature can turn any drawing into a full-color image, complete with the designer's art style of choice, including photorealism.

Templates Based on Mood Boards or Prompts

One of the most intriguing features in development is the ability to create templates based on mood boards or prompts. Designers submit their chosen text prompt into Firefly, and the program generates intricate templates. Having a mood board as the basis of your design means that artists can automatically generate on-brand material for their clients.

Personalized Results

Adobe's Firefly AI can remember several images uploaded by the user as an option to include in generated art. This allows users and designers to include their own products and brands, giving them a chance to showcase them in the generated art. This feature can create brand consistency and save designers valuable time.

3D-to-image

Artists can use a 3D model as inspiration for images, and they can apply various textures to completely change the appearance of the object. You can select predefined segments of your model to apply different textures to, making it easy to get the exact effect you’re looking for.

Generating Color Grades

One particularly exciting feature that is currently unconfirmed is the ability to generate color grades. This feature allows artists to enhance or alter the colors in their work to achieve a certain look, mood, or style. You can apply it to the entire scene or just change the color grades of a few elements, such as illuminating faces. The possibilities for creating impactful and visually stunning work are endless. From generating major changes to your scenes, including changing the seasons and weather, all with the simple writing of a prompt.

Find B-Roll

Finally, another unconfirmed feature that has been hinted at is Firefly’s ability to find B-roll. Using natural language to dictate a story, the AI can scour through video to find scenes containing the elements, characters, or actions needed. This feature has enormous potential for those working with video, making it easier and faster to find the right footage and create compelling content.

In Closing

The Adobe lineup makes image-generative AIs look like toys. As spectacular as MidJourney, DALL-E, and Stability may look today, they are not professional tools yet. A design tool for professionals allows a lot of adjusting and customizing.

Firefly doesn’t promise a generative AI for wholesale images, it proposes individual tools that respond to text and image prompts. This would be unique, as it would allow creators to make the minute changes that they need to their images. These are tools dedicated to professionals instead of fun and spectacular toys. They set a target for what can be achieved in the field using AI.