Adobe is significantly enhancing its experimental iOS camera application, Project Indigo, which initially launched last year. Known for its professional controls, multi-frame super-resolution, and diverse capture modes, the app is now integrating advanced AI capabilities. These new features leverage large language models (LLMs) to provide insightful critiques of photographs and offer tailored editing suggestions.
Beyond LLM-driven feedback, Project Indigo is also introducing other compelling AI functionalities, including sophisticated object removal, the ability to generate depth of field for images, and a range of stylistic transformations for photos.
Marc Levoy, the visionary leading Adobe's Project Indigo, previously played a pivotal role in developing the acclaimed camera technology for Pixel phones. Levoy highlighted that many existing generative AI tools rely on prompt-based editing, a process that can often be challenging when trying to achieve a precise or desired outcome. To address this, most of Project Indigo's new experimental AI features are designed as straightforward buttons, engineered to produce more predictable and deterministic results.
For context, Google's camera coach feature for Pixel phones, introduced last year, offered more general framing advice. In contrast, Project Indigo's features are notably more descriptive, providing users with specific guidance that could genuinely enhance their photography skills, even if they occasionally diverge from the AI's assessment.
Within this category of descriptive AI guidance, two primary features stand out. The first is the photo critique, which delivers a "professional" assessment covering crucial elements such as framing, lighting, color composition, and the overall emotional impact of the image.
The second feature offers capture and edit suggestions, providing actionable tips for reshooting a photo. This includes advice on adjusting framing, exposure settings, and even manipulating objects within the viewfinder. For example, the app once advised me to remove a hexagonal white object from a photo I had taken.
Furthermore, a dedicated section within the app guides users on how to improve an existing photograph by utilizing the powerful controls available in Adobe Lightroom.
While object removal has been a staple in popular photography apps like Apple Photos, Google Photos, and Adobe Photoshop for years, these tools typically require users to manually circle or draw around an object on the screen, a method that frequently falls short of perfection.
Project Indigo's innovative object removal feature streamlines this process by offering intuitive toggles for common elements to be eliminated from an image. These include people in the background, trash and refuse bins, wires and poles, fences, vehicles, and general clutter. Users can also precisely describe a custom object for removal, with the feature yielding remarkably impressive results.
In one instance, the app flawlessly removed both my friend and an object he was holding from the background of a photo, without introducing any noticeable artifacts or distortions.
The app also harnesses AI to intelligently create depth of field in photos, effectively simulating a professional-grade blurred background.
With this latest update, Adobe is also exploring a style transfer feature, allowing users to transform their images into various artistic renditions such as watercolor, pen and ink, ink line with color wash, monochromatic, or a backlit subject effect. While some of these styles evoke memories of the early days of the Prisma app, and the outputs are refined thanks to advanced models, the author notes that style transfer itself is not a novel or particularly groundbreaking feature.
Interestingly, despite Levoy's earlier critique of prompt-based AI, Project Indigo does incorporate a feature where users can describe their desired edits. This enables the execution of complex editing workflows not available through standard preset tools. However, the author cautions that this flexibility could potentially lead to less curated or "sloppy" results.
Underpinning these new functionalities, Adobe is utilizing Google’s Gemini-based Nano Banana model, though the company remains open to integrating other models, including its proprietary Adobe Firefly. These features, housed within an "AI playground" tab, are currently in a testing phase, accessible only to a select group of users. While there's no guarantee these capabilities will reach a wider audience, the author expresses optimism, viewing them as valuable tools for fostering an understanding of photography's nuances, rather than merely contributing to an influx of uninspired content.
The Editorial Staff at AIChief is a team of professional content writers with extensive experience in AI and marketing. Founded in 2025, AIChief has quickly grown into the largest free AI resource hub in the industry.