Yesterday April Fools Day and then Sam Altman announced 4o image generation for free.
But it's actually only available a limited number of times per day, and free users have to wait a long time each time.
even ifGPT-4oIt's unavoidable that the image generation is free and still many people can't use it because of the internet.
There are also a lot of people asking what to do if you can't use GPT-4o. You can tryBean curd. While it is true that many of the current features of the beanbag are not yet as powerful as the GPT-4o, they are sufficient for most people, after all, the vast majority of people are merely putting together a game.
Today, we'll take a good look at Beanbag's seven features for image processing.
I. Image Generation
Since today's talk is aboutDoubao AIof image features, that first one is definitely image generation.
We open the beanbag, click on Image Generation on the left, output the prompt word on the right, and below that we can choose the scale and style we want.
Click Generate and four images will be generated.
Take a look at the effect of the generated image:
Below we have chosen one of them to demonstrate the other six image manipulation features of Beanbag.
Second, the picture to draw
First, let's talk about image transcription, which is really a style transformation of an image.
We can see that GPT-4o's recently produced Ghibli-style pictures have exploded all over the Internet, but this kind of picture style transformation has actually always been just the most basic function of AI painting.
We click on the reference image and upload the image we just generated. Then enter the prompt word:Change the image to read: Ghibli Style.
III. Keying
Click on Keying at the bottom of the dialog box. Then upload the image and click on Keying out the subject.
A picture with a blank background did the trick.
IV. Erasure
Erase refers to the ability to remove unwanted elements from a picture.
For example, click on the Erase function and then apply the font on the clothes. We can then clear remove the font from the clothes with one click.
V. Expansion of the map
You can also expand the image with one click to make more content appear in the image.
VI. Partial repainting
Local repainting means, taking a part of the image and re-generating it based on the cue word.
Scribble a little, then type in the cue word, a phrase to get the girl to put on her sunglasses.
VII. Intelligent Editing
Intelligent editing, and one of the most powerful features of the current beanbag. It can basically include each of the above functions.
Just one sentence is needed to change the map. And it supports continuous chart change.
For example: changing the background
Another example: wearing sunglasses
Another example: changing clothes to blue
It's easy to change the image with just one sentence, and it's very powerful.
Have you learned all seven functions of the beanbag above? Learn them and use them. Don't get hung up on not being able to use the GPT-4o anymore.
I'm sure Beanbag's next update won't be worse than GPT-4o.