Google's Circle to Search Now Identifies Multiple Objects in a Single Image at Once
Summary
Google's Circle to Search and Lens now use a powerful 'fan-out' technique driven by Gemini AI to simultaneously identify and search multiple objects in a single image, delivering one cohesive response in seconds for complex visual queries about outfits, rooms, and more.
Key Points
- Google's Circle to Search and Lens now use a 'fan-out' technique powered by advanced Gemini models to simultaneously search for multiple objects within a single image, delivering one cohesive response instead of requiring item-by-item searches.
- AI Mode acts as the 'brain' that performs multi-object reasoning on images while the visual search backend serves as the 'library,' triggering dozens of parallel searches in seconds to answer complex visual queries about entire scenes, outfits, or rooms.
- Users can also initiate the fan-out process from text searches in AI Mode, allowing them to refine results conversationally by referencing specific images and expanding visual searches beyond shopping to use cases like museum art explanations or identifying bakery items.