At this week's ActiveRain Zoom call, Praful Thakkar provided an interesting overview of an AI graphics package that uses a tool called NanoBanana. This tool is accessed through Google Gemini, and it allows you to take a photo and make changes to the photo to improve it.
So first, we were instructed to go to https://gemini.google.com/app. I tried four different photos and was not charged for usage, so I don't know if this is always free or whether there will eventually be a charge. But it was totally free to try.
Once on the website, a photo is uploaded by clicking on the + sign and selecting the photo that is to be changed. Once it's uploaded, the user can tell the system how they want to change the photo.
So for my experiment, I uploaded this photo:
As you can see, it's an autumn photo and many of the trees have no leaves, and there are some branches in the way of the home. So my prompt was as follows:
Please make the following changes to this photo: remove the leaves on the ground and make the grass a dark green. Make the bushes in front of the home a healthy green and make the bushes sculptured. Also remove the tree branches that are in the front of the photo in front of the house.
After it did that, I added this prompt "Also, please make the trees behind the home full of green leaves". And the final picture was as follows:
Now the program didn't totally follow instructions. You can see there are still some leaves on the ground but otherwise, this was a good result from a quick prompt. The entire experiment took about 4 minutes to complete.
Bottom line, I would use this tool again. The trick to getting good output is giving the right prompts, so I'm sure my prompts could use some work!
And that's my homework for this week's Zoom call! Come join us next Tuesday!
Comments(22)