Google Photos has introduced a new "Auto frame" feature powered by machine learning and generative AI, allowing users to re-compose existing photos from new perspectives. The system interprets images as 3D scenes, enabling adjustments to camera position and focal length to reveal previously unseen content. This approach involves a two-stage process: first, estimating the 3D scene and camera parameters, and second, using a generative latent diffusion model to fill in any newly revealed areas. AI
IMPACT Enhances user experience in photo editing by enabling novel re-composition of existing images.
RANK_REASON This is a product feature update using AI, not a core AI model release or research breakthrough from a frontier lab.
Read on Google AI / Research →
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →
