In Adobe's latest version of Photoshop (Beta), a new AI image editing function (function name: generative fill) has been added. According to the official introduction, generative fill only needs to be described by text or without description. Use AI tools to delete elements, add elements, expand pictures, and other operations on pictures in Photoshop.
Test version: PS 2024 Beta Mac25.0 version
To use the Photoshop (Beta) version, we need to download and install it in Creative Cloud. First, let’s install Adobe Creative Cloud. The Adobe account can be registered with a google account (the Photoshop (Beta) version may not be available in non-US regions).
After the installation is complete, click "Beta Application" in Creative Cloud, you can see the Photoshop (Beta) version, click "Try", the program will be installed automatically.
The installation progress percentage will be displayed here, wait for the installation to complete, this installation process is still relatively long, you need to wait for a while.
After the installation is complete, open Photoshop (Beta), and click "Open" to import a picture.
First of all, let's try the picture completion function, use the "cropping tool" function to enlarge the width of the picture canvas.
Then use the "Selection Tool" to select the area that needs to be expanded. At this time, you can see the "Generative Fill" button below.
After clicking "Generative Fill", we need to enter the prompt word or leave it blank. Here we need to note that the prompt word currently only supports English input, and the program will automatically match it if it is left blank. Here we leave blank and click "Generate".
Photoshop will automatically start calculating the image content that needs to be completed. It should be noted that factors such as the size of the image and the network will also affect the duration of the final calculation.
OK, we got the final result. After computing the result, there are 3 pictures to choose from on the right side of Photoshop (AI will generate 3 schemes each time).
Judging from the final AI calculation effect, it may basically match the picture in terms of style and style, but the processing of details is relatively ordinary, especially in the processing of areas that are handed over and fused with the original picture.
As you can see in the layers, a new "generative layer" has been created, so the original image has not been changed, which is also a major advantage in Photoshop.
Next, try the function of removing elements in the picture. We use the "Lasso Tool" to select the two people on the right, click "Generative Fill", enter "delete character" and click "Generate".
After waiting for the calculation to finish, you can see from the resulting image that the characters have been removed very well, and the splicing of the ground texture is also similarly good. Although this kind of operation of removing characters can also be done by conventional PS, the overall operation cost and difficulty are greatly reduced after using AI.
Finally, let's try to add elements to the picture. Here we want to add "moon" to the picture. Still use the selection function, click "Generative Fill", enter "moon", and click "Generate".
Judging from the generated results, the task has indeed been completed, but judging from the overall fusion effect of the picture, the newly generated moon is not in harmony with the overall picture, and the degree of fusion is poor. The example picture below has been selected to look the best one of them.
