We’re launching Lyria 3.5 in Google Flow Music, with advances across musicality, lyrics, vocals, and creative control
deepmind.google·6d ago
AI Engineering
DeepMind Blog·Oct 23·5 min read
Image editing in Gemini just got a major upgrade
TL;DR
Previously, editing images of people and pets often resulted in inconsistent likenesses. The new image editing model in the Gemini app from Google DeepMind allows users to change outfits, blend photos, and apply styles while maintaining a consistent appearance.
✦ Why It Matters
Engineers can leverage this model to enhance user experience in image editing applications by ensuring consistent subject likeness.
Key Takeaways
- Upload photos to maintain likeness while editing.
- Experiment with blending multiple images for unique scenes.
- Use multi-turn editing for iterative changes to images.
- Apply design elements from one image to another seamlessly.
Full Summary
The Gemini app has introduced a new image editing model, enhancing its native editing features. This upgrade focuses on preserving the likeness of individuals and pets during edits, addressing the common issue of subtle inaccuracies in edited images.
Users can upload photos and specify changes, such as trying on different outfits or placing themselves in new backgrounds, while ensuring they still look like themselves. The model also allows for blending multiple images to create unique scenes and supports multi-turn editing, enabling users to make iterative changes.
Additionally, users can apply design elements from one image to another, such as textures or colors. This update aims to provide users with more creative control and flexibility in their image editing endeavors.