Last night, ByteDance's JiMeng AI launched the Image 2.1 model, which solves the problem of AI-generated images not being able to produce Chinese fonts. In the future, creating posters will only require a simple text description, significantly improving the efficiency of designers.
The usage is also very simple:
Open the JiMeng official website: https://top.aibase.com/tool/jimeng
After entering the image generation page, simply input your description in the text box.
Detail the elements you want in the image, as well as the text and font you wish to include.
For example, if you input in AIbase: "A poster for a historical drama featuring a eunuch from the Qing Dynasty, with Zhao Mengfu's cursive writing saying 'Reborn as a Eunuch in the Qing Dynasty.'
Then I can obtain a poster like this with text. As you can see, the main text part is already very perfect, while the smaller text may still have some garbled characters, but this is already a historic improvement.
You can also use this model to generate red envelope cover images.
For example, input: "LOGO, great fortune, Guochao style, red and gold festive, Yan Zhenqing's regular script 'Great Fortune,' serpent totem."
For this type of material image, with fewer additional elements, the text representation needs to be better.
Let's try directly inputting an image of a capybara and creating a logo:
Now, let's create a movie poster:
Prompt: "A highly cinematic sci-fi poster, with a massive spaceship exploding and disintegrating in space, and a background of a shattered Earth. The title is 'Doomsday Approaches,' with the top text reading 'Releasing on New Year's Eve 2026,' and the bottom credits stating 'Director: Wong Kar-wai, Produced by: AIBASE.' The image should be vibrant, dramatic, high contrast, in 8K quality, with cinematic lighting effects."
What do you think? Does it have that feel? Overall, after testing, the new model is quite fun to use, and everyone can freely unleash their imagination to create the text posters they desire.
This model has the following application scenarios:
Personalized poster creation: Creators can freely input themes and desired fonts to generate customized movie or event posters.
Novel and book cover design: Create attractive covers using only text descriptions.
Comic scene creation: Accurately recreate characters and atmospheres.
Social media visuals: Add visual appeal to content.
Story creation assistance: Inspire creativity through images.
Virtual character design: Generate unique images with simple descriptions.
Holiday card creation: Personalized blessings at your fingertips.
Game interface elements: Provide creative materials for games.
The model is currently still in a beta testing phase, so those who haven't used it yet can wait a bit longer; it is expected to be fully launched soon.