The popular PS AI expansion. Make your own memes
In the past two days, a group of AI-enhanced pictures of film and television emoticons have become popular on the Internet.
The ambitious "Oboi" who looks like "I want it all" is actually stepping on red high-heeled shoes and being embraced by Master Xing:
Brother Xueyou's famous "eating shit, you" emoji was originally said to a pot. I don't know if there is still Jiuzhuan large intestine in the pot:
In fact, such a meme is not difficult to make, as long as you have an internet connection and a computer, you can do it too. It doesn't matter, today we "teach a man how to fish rather than give him a fish", and teach you how to use AI to expand the picture.
Step 1: Download Creative Cloud
After all, just one sentence:
Install the overseas version of Photoshop (Beta), and use the Generative Fill tool to expand the image with one click.
We open the download link of the Creative Cloud installation package of the overseas Beta channel, and the browser will automatically download the installation package:
If you are a Windows user , double-click the downloaded installation package, and the installation program will automatically open the browser to proceed to the second step.
If you are a Mac user , open the downloaded DMG image file, double-click the Creative Cloud Installer program, and click Open in the pop-up warning window:
The installation program will start, we click to continue , enter the system password in the pop-up dialog box, click OK :
Step 2: Register an account and install Creative Cloud
Next, the installer will prompt us to create an account. If you have a Google/Apple account, you can log in with it. We recommend creating an account :
Enter the email address, password, name and other information. For the country/region, please select Hong Kong Special Administrative Region of China , and click Finish .
If the webpage prompts you to log in from a remote location, click Approve Login .
Next, the installation package may prompt to install the genuine service AGS, we skip it and click Continue , wait for the installation package to complete, and Creative Cloud will open automatically.
Mac users, please note that after the installation is complete, the Mac computer will prompt "Open Login Items". If you just want to experience AI image expansion, you can choose No, thank you here.
Step Three: Install Photoshop (Beta)
At this time, we can successfully enter the main interface of Creative Cloud. If there is a pop-up advertisement, just close it in the upper right corner.
We choose Beta application → Photoshop (Beta) → trial , and then wait for Photoshop (Beta) to be downloaded and installed automatically, so that we can try Photoshop for free for 7 days, which is generally enough to enjoy.
After the installation is complete, click Start Trial to open the Photoshop (Beta) program. If you want to use it later, we can find the icon of Photoshop (Beta) through the Windows Start Menu/Mac Launcher and open it directly.
We close the "New Features" introduction in the main window of Photoshop (Beta), click Open →import pictures on your computer , or directly drag and drop pictures into the main window.
For Mac users, please note that if a file access warning pops up when opening the file, just select OK .
Photoshop will open the picture, and pop up the relevant tutorial of "Generative Filling". Interested friends can read it. If you just want to try AI image expansion, you can close the tutorial pop-up window and continue with our subsequent tutorials.
Step 4: Practice AI image expansion
The film and television emoticons we practiced this time are taken from "The Romance of the Three Kingdoms", "I have never seen such a brazen person." This emoticon pack can be said to be a long-lasting ghost material. Let's see if it can be played in the hands of AI What's new:
Here we also provide the original picture, you can save it for easy operation and learning:
Since we want to expand on the basis of the original image, the first step is to select the cropping tool on the toolbar, and there will be 9 adjustable anchor points around the image. Press and hold the mouse outside the anchor point according to our needs. Just drag and go:
You can expand the canvas to such a degree by imitating the picture below, and then press the Enter key / double-click the picture itself to confirm the operation.
Then we choose the Rectangular Marquee Tool to frame the content of the picture that is a little smaller than the original picture. The reason why the original image is not strictly selected here is to allow AI to refer to this small part of the content we left behind, and better generate an expanded image that can blend with the atmosphere of the original image.
The next step is the most important step. We need to click Inverse Selection from the Select menu on the menu bar to switch the selection to a part of the canvas other than the original image. How can I confirm that the reverse election was successful? You can see if there is a flashing dotted line on the outer edge of the canvas, if there is, it means that the "reverse selection" is successful:
Now, let's click the Generic Fill button below to enter the AI image expansion interface:
Here, we can directly click Generate to let AI automatically expand the image; we can also enter the prompt words you want (English only) and then click Generate to let AI generate the image expansion effect you need.
Practice 1: AI automatic map expansion
Let's try the effect of directly letting AI automatically expand the image. Click directly to generate , and you can see that AI only draws the whole body of the prime minister, which is quite satisfactory and conservative. Of course, for a productivity tool, stable output is sometimes an advantage.
We can let AI generate more pictures by clicking the Generate button, or browse the generated picture results by switching < and > next to the button.
But having said that, since what we are doing is the AI expansion of memes, how can we not do something?
Practice 2: AI image expansion with prompt words
Then let's try the generation effect of prompt words again. Enter "Ride a motobike" (riding a motorcycle) in the added prompt on the left. Riding a motorcycle in ancient China seems absurd but also a bit fun. Sure enough, AI understood their intentions and generated a set of somewhat unconstrained expansion pictures:
We can still adjust the prompt words, browse the generated images, and continue to generate new images through the quick generation bar below the image, or view the generation effect through the "Change" panel on the right sidebar.
Step 5: Export the image generated by AI
After generating and selecting the AI expansion image we want, we can select File → Export → Quick Export to PNG in the menu bar, and then we can rename and save the generated image.
It can be seen that the expanded image generated by AI is still not stable for the depiction of human hands. We can obtain better images through multiple generation and post-cropping. .
In the process of writing this basic tutorial, we also encountered some program problems. You may also encounter them in actual operation. Here we share our solutions for your reference. If you have more questions, please comment and leave a message, and we will try our best to help you solve them.
Easter Eggs: Frequently Asked Questions
1. During the generation process, I got a prompt of "The temporary storage disk is full", and I cannot expand the image with AI.
This is because there is not enough free space on the hard disk where the Photoshop temporary disk is located. Photoshop will use the system disk (Windows is usually the C disk, and Mac can be directly understood as the hard disk) as the temporary disk. Here are two solutions:
A. Clean up the files in the partition and ensure that there is a certain amount of free space under the system disk
B. Hold down the Ctrl + Alt keys ( Windows ) or Cmd + Option keys ( macOS ) during Photoshop startup, then set up a new scratch disk for the system under the Scratch Disk Preferences dialog that opens.
2. Photoshop prompts "The generated images have been removed because they violate the user guidelines", and the AI can no longer draw.
This type of problem is generally caused by the original image itself triggering the review standard of the Adobe server. Some violations we understand conventionally, some illegal text in some subtitles, and edge-rubbing prompts may trigger such problems.
In this case, you can try to remove the problematic parts first, such as body parts that may violate regulations, subtitles and prompt words that may trigger review, etc., and then try again. If it doesn't work, I can only ask everyone to change to a more "healthy" original picture and challenge Adobe again.
3. Photoshop prompts "We are facing peak demand. Please wait a moment and try again as soon as possible. Thank you for your patience", AI stopped drawing.
Generally, it is during the peak period of the server, because the function of AI image expansion requires Adobe server cloud computing, and such prompts will appear when too many people use it at the same time.
We actually measured that 10:00 am to 12:00 am Beijing time is the off-peak period, and the use is relatively smooth.
I hope this zero-based PS AI image expansion tutorial from installation to practice can help you make a meme that you are satisfied with. Welcome to follow us, and there will be more AI contents delivered to you in the future.