Adobe’s Project Indigo introduces AI photo critique and deterministic editing tools
The latest update to Adobe’s Project Indigo leverages Google’s Gemini-based Nano Banana model to offer professional-grade photo analysis and precise background clutter removal, currently available to a select group of testers.
Adobe has rolled out a significant update to its experimental iOS camera application, Project Indigo, introducing a suite of artificial intelligence features designed to move beyond traditional prompt-based editing. The update, which is currently in a testing phase for select users, integrates large language models to provide structured, deterministic outputs through specific interface buttons rather than relying solely on open-ended text commands.
At the core of the update is an AI-driven photo critique tool that offers professional feedback on framing, lighting, colour, and emotional impact. Alongside this analytical capability, the app provides real-time capture and edit suggestions, offering users tips on how to improve shots by adjusting exposure, framing, or removing specific objects from the viewfinder. This approach aims to educate users on photographic nuances, distinguishing it from more generic coaching features previously seen in competitor devices.
The update also addresses the common challenge of background clutter with advanced object removal capabilities. Unlike manual selection methods used in apps such as Apple Photos and Adobe Photoshop, Project Indigo allows users to toggle the removal of specific clutter types, including people, vehicles, wires, and trash. Users can also describe custom objects for removal, with initial tests indicating the tool can eliminate subjects without creating visual artifacts.
Additional editing features include depth of field generation to simulate blurred backgrounds and style transfer effects that convert images into tones such as watercolour, pen and ink, and monochromatic styles. While style transfer is not a novel concept in photography software, the integration here is powered by Google’s Gemini-based Nano Banana model. The application also retains an "AI playground" tab for descriptive edits, with Adobe considering the inclusion of its own Firefly model as an alternative engine.
Marc Levoy, who heads the Project Indigo project and previously developed camera capabilities for Google Pixel, has positioned these features as a response to the difficulties users face in finding the perfect prompts for generative AI tools. By offering more descriptive and actionable feedback, the update seeks to provide practical utility for photographers. The features remain exclusive to a testing group, with no confirmed timeline for a wider public release.
