Adobe Tests On-Device AI Photo Editing and Guidance in Project Indigo App
Adobe has released version 1.1 of its experimental Project Indigo camera app, introducing generative AI features powered by Google's Nano model. These features, housed in the "AI Playground," offer real-time shooting advice, custom prompt editing, and on-device photo adjustments like object removal and lighting tweaks. This update highlights the growing trend of edge AI, bringing complex generative photo editing and real-time composition guidance directly to mobile devices without relying on cloud servers. It showcases how computational photography experts are pushing the boundaries of mobile camera applications. The AI Playground feature is currently in a limited, free trial phase and may eventually become a paid service. To ensure user privacy, all processing is done on-device, and Adobe does not upload photos, prompts, or user data to its servers or use them for model training.
## BACKGROUND
Project Indigo is an experimental computational photography app developed by Adobe Labs, led by former Google Pixel camera experts Marc Levoy and Florian Kainz. Google's Gemini Nano is a lightweight AI model designed specifically for efficient, on-device tasks on mobile hardware, enabling fast processing without internet connectivity.