A few days agoOpenAI release ChatGPT OpenAI has launched a new feature, DALL•E 3, in ChatGPT.Image Editinginterface,The image can be partially modified.
It has to be said that OpenAI is good at converting traffic.
Although this feature is not particularly eye-catching, it has a certain entertainment value and can well capture the user's desire to experience.
However, this feature is not free and you need to open a Plus account to experience it.
This wave of operations can be said to have directly dispelled the concerns of a group of people a few days ago about whether OpenAI would incur a loss.
By the way, how is DALL•E 3?
Let me first state the conclusion.It's quite interesting, and it can also be used to generate traffic and monetize.
To know how to do it, let us first understand what the content of DALL•E 3 is after the update.
Input prompt: Help me draw a puppy playing on the grass.
After you select the picture you like and click on it, the following editable interface will pop up. A series of options are provided in the upper right corner of the interface. Click the Select button in the green box to proceed to the next step of editing.
As shown in the picture, on the left is a draggable button for selecting the appropriate size of the smear tool. In the lower right corner, you can modify the area you have smeared through text prompts.
For example, my hint is that there is a lion lying in the distance ready to catch a puppy. The effect is as follows:
At this moment the puppy thought in his heart: Okay, okay, if you want my life just say it!
However, DALL•E 3 can not only add details,Different style options have also been added to make the drawing style more diverse.
In addition, DALL•E 3 can also perform detail removal adjustments and detail regeneration.Perform detail removal and detail regeneration.
Here I chose an anime style to draw a picture to show everyone how to delete, adjust and regenerate details.
Take this picture as an example. I am not very satisfied with the entire scene and want to change some details. For example, I want to delete this seagull. How should I do it?
Same as above, select first, adjust the size of the tool, then fill in the prompt words: remove this seagull.
The effect is as follows:
How about it, isn't it completely consistent, right?
So how do you regenerate the details?
Similarly, first select the target area, then input the prompt word for the desired modification, for example: turn this land into clouds.
Effect image:
I discovered a little trick during my use. In addition to using the selection tool, users can also directly enter editing requirements in the dialogue panel on the right side of the editor interface without using the selection tool, except for this case, which will most likely change the style and style of the original picture.
However, this will most likely change the style and style of the original picture, so I won’t explain it here.
DLL•E 3 has a significant advantage over MJ and GPT text-to-image models, that isIt can ensure the consistency of the image by extracting the seed value.
The children's picture book market is still quite popular at present. Partners who have the conditions can use the painting function of DALL•E 3 to create children's picture books and videos, and sell collections.
If you want to go a step further, you can publish a tutorial on how to use DALL•E 3 to create children's picture books.
Well, the meal is ready. Whether you eat it or not is up to you.