Project Indigo Gains AI-Powered Photo Critique and Editing Suggestions

Adobe is expanding its experimental iOS camera app, Project Indigo, with a suite of new AI-driven features designed to critique photos and provide actionable editing suggestions. Launched last year, the app originally offered pro-level controls, multi-frame super-resolution, and various capture modes. Now, it integrates large language models (LLMs) to deliver a 'professional' opinion on aspects like framing, lighting, colors, and emotional impact. The feature, led by Marc Levoy, who previously developed Google Pixel's camera technology, aims to move beyond generic suggestions. According to Levoy, most generative AI tools rely on prompt-based editing, which can be tricky for users. Instead, Project Indigo's new features use buttons for more deterministic outputs, making them more accessible. The app can also offer capture and editing suggestions, such as advising users to remove an object from the frame or adjust exposure.

Advanced Object Removal with Toggle-Based Controls

Photography apps like Apple Photos, Google Photos, and Adobe Photoshop have offered object removal for years, but they typically require users to manually circle or select objects. Project Indigo's new feature simplifies this by providing toggles for common distractions, including people in the background, trash and trash cans, wires and poles, fences, vehicles, and other clutter. Users can also describe a custom object to remove via text. In tests, the feature performed impressively, removing a person and an object from the background without creating noticeable artifacts. This approach reduces the friction of precise manual selection, making the tool more efficient for quick edits. The feature is part of a broader push to integrate AI into photography workflows, aiming to save time and improve results for both amateur and experienced photographers.

Depth of Field Simulation and Style Transfer Features

Project Indigo also introduces AI-powered depth of field simulation, which creates a blurred background effect similar to a portrait mode. This allows users to add a professional look to photos taken in less-than-ideal conditions. Additionally, a style transfer feature lets users transform their images into various artistic tones, including watercolor, pen and ink, ink line with color wash, monochromatic, and backlit subject. While style transfer is not a new concept—apps like Prisma popularized it years ago—the outputs here are more refined due to advanced models. However, the feature's utility is debatable, as it may not offer significant value beyond novelty. The depth of field simulation, on the other hand, could be more practical for users seeking to enhance portraits or isolate subjects.

Powered by Google Gemini Nano, With Potential for Adobe Firefly

Adobe is using Google's Gemini-based Nano Banana model to power these new AI features, but the company remains open to swapping in other models, including its own Adobe Firefly model. This flexibility allows Adobe to optimize performance and capabilities over time. The features are bundled under an 'AI playground' tab and are currently in a testing phase, with access limited to select users. This cautious rollout suggests that some features may not reach a wider audience if they prove unstable or unpopular. Despite Levoy's criticism of prompt-based editing, Project Indigo includes a prompt-based feature that lets users describe edits not available in preset tools. This could lead to unpredictable results, but it offers creative flexibility. The use of on-device AI like Gemini Nano also ensures privacy and faster processing.

Implications for Photography Education and the Risk of 'Slop'

Project Indigo's new AI features have the potential to educate users about photography nuances, such as framing, lighting, and composition, rather than just generating automated edits. The photo critique feature, in particular, offers descriptive feedback that can help users improve their skills, even if they disagree with the AI's assessment. However, the prompt-based editing feature, despite Levoy's criticism, introduces a risk of producing 'slop'—low-quality or unintended results. This dual nature highlights the challenge of balancing creative freedom with quality control. As AI tools become more integrated into photography, companies like Adobe must ensure that features enhance, rather than replace, human creativity. The limited testing phase suggests Adobe is cautious about this balance, but the potential for misuse remains.