Image processing in Gemini
Google is betting full on AI tool Gemini. The application will be enhanced with new photo editing and image processing capabilities. The updated version should make sure that when you edit a photo, you or someone else looks the same as the original photo. You should see this in the new version of Gemini. It does not matter if you choose a different background, or choose a different outfit. This applies not only to people, but also to pets you include in the edit.
From now on, the Gemini app gives you several new options. We highlight those below;
- Costume or location change: you can put yourself in a different outfit, or place yourself in a different location; without changing the look.
- You can combine photos: for example, by merging a photo of yourself with that of your pet.
- Multi-turn editing: this feature lets you edit a photo step-by-step. For example, you can add furniture to a room or adjust specific elements while everything else remains the same.
- Combine designs: you can use a pattern or texture from one image, on an object in another photo. Think, for example, of petals on rain boots.
You can use the final result in a new creation in Gemini, or turn it into a short video, for example. Google emphasises that edited images are given both a visible, and an invisible (SynthID) watermark. That way, it is clear that the images were created by AI.
The new photo edits can be used in Gemini from now on. From our brief test with a car, the system seems to do a pretty good job. Even shadows and details match pretty well, like the licence plate. However, Gemini couldn't handle the precise rims very well with this photo, but the other details are surprisingly good!
