Nano Banana!

Something a little different today: nano banana!

Or what I often like to say to myself since the term is kind of funny: nana banana nana banana nana banana.

Nano banana is a new image creating tool within Google’s Gemini generative AI. I played around with it and found it to be remarkably good. I think it may be the best image maker outside of apps such as MidJourney, Adobe Firefly, Dall-e 3, Stable Diffusion, Leonardo, Bing Image Creator.

Gemini’s version is FREE.

1.a Original Picture

Here’s the original image that I sourced from Pixabay.

1.b Without Background

First thing I did was to ask it to remove the background: “In the uploaded image, can you remove the background so that only the woman remains?”

The result:

Pretty good, huh? Note though the background is white, not transparent.

1.b Glowing outline

The next experiment was to add a glowing outline to the woman. This attempt was due to a question someone asked me on how I was outlining myself in my pictures. I was trying to explain to the person who I did it, but it required certain tools. I thought nano banana might be the way for this person to go.

The prompt was: “Wow. Let’s go back to the original image with the background. Can you apply a kind of neon orange glowing light around the woman but keep the background. Just apply the light.”

I think this worked very well. You might be able to get various outlines depending on the choice of words. You might request an inky outline or something with a diffused light.

1.c Change the Scene

I wanted to explore changing the background, keeping in mind the person who requested the help. I’m thinking that person might want to change his background into something closer to what he is working on.

The prompt was: “In the image with the background and the glowing orange line around the woman, can you change the background to a scene of a server room in the dark but with ambient lighting, giving a futuristic feel?”

In this output, it looks like Gemini also changed how the light hit the woman. The light hitting the woman was adjusted to fit the darkened room atmosphere.

So, far it is looking good!

1.d Transparent Background

My last experiment was this particular image was to see if I could get a transparent background from Nano Banana.

Prompt: “Go back to the original image. This time I want you to remove the background but instead of a white background, I need it to be transparent. Can you do that?”

Okay, that didn’t work. Notice the botched “transparent” background to the upper right and to the right of her head? Yeah, that is not a real transparent background.

If you ever wanted to cut out the lady to merge with other pictures in another program, I would first ask to get rid of the background – replace with white background maybe – and then use the other program to cut out the white.

Or if you need to merge pictures…well, Gemini can do that too.

2.a 3 images to merge

All three images come from Pixabay and I want to merge the food items into one image.

2.b Merged Image

Here’s the prompt I used to merge the three food items into one image: “Can you combine the hamburger, the fries and the milk shake from the three images into one, with a background of early evening light fading away after a sunny day.”

And out popped the following image:

That actually looks pretty good to my untrained eye. The food items appear to sit on the table with proper perspectives and consistent shadow directions. And while the background has an interesting blurring technique, I’m not opposed to it.

In a way, the blurring technique adds to the picture.

3.a Original Images

My last experiment will be to change dress styles and colors. This test is in preparation for when I might have to change what I’m wearing in my pictures. You know, take me out of bathing suits into business suits.

3.b Catwoman

The first test I want to do is to transform the woman in the black turtleneck into catwoman.

The prompt is as follows: “Hello, can you transform this woman into one wearing a cat suit, like the superhero Catwoman? And change the background into a New York night scene but darkened and without a lot of people.”

I had to do a correction because Gemini only did the background. So the second prompt became: “Okay, keep the background but put her in Catwoman outfit, kind of like superhero Catwoman. Maybe replace the dark glasses with that catwoman mask.”

Finally, I get:

I don’t know about you, but I think that is actually pretty good.

3.c Elizabethan Woman

My last test was to transform the nearly naked red head into a person wearing a dress right out of the 17th century.

The prompt was as follows: “Okay, this woman needs to be transformed into a bright red head – more like a fiery red-head. And change her clothes into something more 17th century garb of a wealthy woman, not a poor woman.”

And ta da!

Now, isn’t that stunning how Gemini managed to keep her face recognizable, even though she is facing in a different direction?

It knew its assignment.

The very last thing I asked was to change the color from blue to purple, and I got this:

4 Conclusions

I think Nano Banana (or Gemini) will suffice for the non-creatives as it is very easy to use and the picture remains consistent with original images. If I want a quick alteration or merging of images, I will be turning to Gemini.