New Gemini Image Model Ensures Consistency in Faces, Pets, and Complex Edits

Google presents an improved image editing model for Gemini, which preserves the similarity of people and pets across various machining steps. All users can access the tool free of charge.

Gemini receives new image editing

Google has introduced a new image editing model from Deepmind for its Gemini app, which on the Images 3 technology is based. The update brings significant improvements to the processing of photos and is available to all Gemini users – not only the subscriber paid versions.

No forgetting characteristic characteristics

The most important progress lies in the improved consistency in the representation of people and pets. The new system ensures that characteristic features such as facial features, hairstyles or fur patterns are reliably preserved even with extensive image manipulations. Users can put themselves or their pets into different scenarios without the original appearance being alienated. An upgrade for image processing in Gemini

Diverse applications

The possible uses are diverse: from changing costumes and places to the combination of several photos to a new composition. Images 3 can create pictures in different styles – such as photo -realistic landscapes, structured oil paintings or bizarre claymation scenes. For example, users can combine a photo of themselves with a picture of their dog and embed both into a completely new environment. The system enables processes in which each change builds on the previous one. This allows empty rooms to be filled with color and then add furniture or other objects. The transmission of design elements between images is also possible – such as the pattern of butterfly wings onto a dress or the texture of petals to rubber boots.

Transparency and labeling

For transparency, Gemini provides every edited image with a visible watermark. In addition, an invisible digital synthid water sign is embedded in order to clearly identify AI-generated content. These measures are intended to help combat the spread of Deepfkes and other manipulated content.

Availability and outlook

The new image generation model is available both in the free version of Gemini and in the advanced, business and enterprise versions. The updated app is rolled out gradually, whereby the complete rollout can take a few weeks. Google plans to continuously expand the image editing functions and integrate other creative tools. The company sees an important building block in AI-based image processing for the future of digital creativity.