Things I want to do
We will use Gemini 2.5 Flash Image (nano-banana) and compare it with Gemini 2.5 Flash Image.
Gemini 2.5 Flash Image (nano-banana) is said to be excellent for image correction.
How to use
Access Google AI Studio.
Please confirm that the area highlighted in red in the upper right corner of the screen is either Gemini 2.5 Flash Image Preview or nano-banana.
If it’s set to something else, click to change it.
Next, enter your prompt in the edit box below and execute. (Pressing Return will not execute it. Click the Run button on the right or press Ctrl+Return.)
Image input
Drag and drop the image into the edit box where you will enter the prompt.
(You can also load data from the + on the right.)
Start a new chat
Click ‘Chat’ on the left side of the screen. (Gemini 2.5 Flash Image tends to try to modify the generated image once it has been created. If you want to generate a new image, it is recommended to create a new chat.)
Compare
Image generation
English prompt
Prompt
A photorealistic close-up portrait of an elderly Japanese ceramicist with
deep, sun-etched wrinkles and a warm, knowing smile. He is carefully
inspecting a freshly glazed tea bowl. The setting is his rustic,
sun-drenched workshop. The scene is illuminated by soft, golden hour light
streaming through a window, highlighting the fine texture of the clay.
Captured with an 85mm portrait lens, resulting in a soft, blurred background
(bokeh). The overall mood is serene and masterful. Vertical portrait
orientation.
translation
A realistic close-up portrait of an elderly Japanese ceramic artist, his face deeply wrinkled and his smile warm and knowledgeable. He is carefully examining a newly fired tea bowl. The setting is his simple, sunlit workshop. The soft evening light streaming through the window highlights the delicate texture of the clay. Shot with an 85mm portrait lens, the background is softly blurred. The overall atmosphere is serene, conveying a sense of masterful skill. A vertical portrait.
Gemini 2.0 Flash Image
Gemini 2.0 Flash Image(nano-banana)
I thought both images were created according to the prompts.
justGemini 2.0 Flash Image(nano-banana)It seems the vertical orientation is being ignored. (This could also be due to the environment rather than the model.)
Japanese prompt
Prompt
A realistic close-up portrait of an elderly Japanese ceramic artist, his face deeply wrinkled and his smile warm and knowledgeable. He is carefully examining a newly fired tea bowl. The setting is his simple, sunlit workshop. The soft evening light streaming through the window highlights the delicate texture of the clay. Shot with an 85mm portrait lens, the background is softly blurred. The overall atmosphere is serene, conveying a sense of masterful skill. A vertical portrait.
Gemini 2.0 Flash Image
Gemini 2.0 Flash Image(nano-banana)
The result was the same even when the prompt was in Japanese.
When creating images in Japanese, simply saying ‘draw an image like XX’ or ‘create an image like XX’ often didn’t work. It was better to use a more detailed prompt like ‘an image like XX’ to ensure image generation was successful.
Logo creation
What I found particularly impressive about this model was the logo creation process.
English prompt
Prompt
Create a modern, minimalist logo for a coffee shop called The Daily Grind.
The text should be in a clean, bold, sans-serif font. The design should
feature a simple, stylized icon of a a coffee bean seamlessly integrated
with the text. The color scheme is black and white.
Please create a modern, minimalist logo for a coffee shop called ‘The Daily Grind’.
Use a clean, bold sans-serif font for the text. The design should seamlessly integrate a stylized coffee bean icon with the text. The color scheme should be black and white.
Gemini 2.0 Flash Image
Gemini 2.0 Flash Image(nano-banana)
I think it’s a matter of personal preference.Gemini 2.0 Flash ImageThen the text is illegibleGemini 2.0 Flash Image(nano-banana)I thought that would be better.
Image correction
I will try to correct images using Gemini 2.5 Flash Image (nano-banana), which is considered to be excellent.
Input image
Change background
Prompt
change background to town
Gemini 2.0 Flash Image
Gemini 2.0 Flash Image(nano-banana)
Change background (Japanese)
Prompt
Change the background to a town.
Gemini 2.0 Flash Image
Gemini 2.0 Flash Image(nano-banana)
Failure (Image was not output.)
In this imageGemini 2.0 Flash ImageWhile it faithfully reproduces the original image, in some cases the resulting image was completely different from the original. (See below)
Also, since things don’t always come across well in Japanese, it’s probably better to do image editing in English.
others
Style Change (Manga Style)
In this exampleGemini 2.0 Flash Image(nano-banana)It’s clear that this version is more faithful to the original artwork.
Gemini 2.0 Flash Image
Gemini 2.0 Flash Image(nano-banana)
Emotional changes (Cry)
In this exampleGemini 2.0 Flash Image(nano-banana)It’s clear that this version is more faithful to the original artwork.
Gemini 2.0 Flash Image
Gemini 2.0 Flash Image(nano-banana)
Add a hat
It may just be a coincidence.Gemini 2.0 Flash Image(nano-banana)A hat that better suits the overall look has been added.
Gemini 2.0 Flash Image
Gemini 2.0 Flash Image(nano-banana)
コメント