Celebrity Photos
This is a new feature of V5.Midjourney In the V5 version, you can know many celebrities. When describing the subject, you can add the name of the celebrity to generate the celebrity's photo. We generate several classic photos from movies:
The first one is Keanu Reeves, and the second one is Vito Corleone (I didn’t enter the lead actor Marlon Brando). The prompts for both sets of pictures are very simple, you can try to write them yourself 😁
The answer is at the end of this chapter.
Clothing Model
Friends who are engaged in clothing e-commerce may need this function:
- Generate a virtual person using AI
- Then send the photo of the goods (such as a piece of clothing) to the AI
- AI fuses virtual people and photos of goods together
Frankly speaking, I think Midjourney should be possible, but the current effect is not very good. To achieve this requirement, you only need to use the blend function mentioned in the previous skill. I will not go into the details here. Basically, I blended the product image and the character image using the blend function:
To verify consistency, I generated several Reeves graphs and tried them separately. The final output was as follows:
Overall, it seems okay, but I think it is unlikely to be used in product description pictures, because the clothes in it are quite different from the original picture (you can take a close look at the buttons inside).
But I think Midjourney may optimize this feature in the future, and I think if you can train the model yourself, this feature should not be difficult to implement.
Old photos
This is a branch of celebrity photos. You may have seen a lot of similar photos recently. The prompt is actually very simple. You can use translation software to adjust the subject and background:
|
Prompt |
explain |
What is the type? |
photography |
photo |
What is the subject? |
a group of Chinese people gathered around Darth Vader |
A group of Chinese people are watching Darth Vader |
What is the environment like? |
on the street |
On the street |
What is the composition like? |
null |
Not important, let AI play freely |
What lens to use? |
fuji film |
To make the photo more interesting, a Fujifilm lens was added. |
What style? |
style of 1990s |
Added a 90s style |
Two sets of images were generated using a similar prompt:
Tip 14: Add style – vintage
For the old photo scene above, we used the 1990s style. Strictly speaking, this 1990s is more suitable to be placed in the subject or the environment. It is also possible to place it in the style, because we not only want the subject to be 1990s but also want the style of the generated image to be 1990s.
This is the last use of the stylistic word: decade.
Take the two prompts in the illustration chapter and modify them. I didn’t add the artist’s name in the prompt, but changed it to style of 1920s. The rose picture has a bit of that style, and the cat picture in the upper right corner is so funny.
Celebrity Photos Answers
Reeves kept it simple, just stating the subject, then noting that he was wearing his signature outfit and walking in the rain:
Keanu Reeves, wearing a black long leather coat, walking down the street in the rain
Then the Godfather's is even simpler:
Vito Corleone sitting on a chair
Tip 15: How to make the people generated by Midjourney more diverse?
The diversity here actually refers to people's height, weight, and even some physical peculiarities.
I don’t know if you have noticed that in most pictures generated by AI, the people are very beautiful and handsome, but this is not good. In fact, it represents social prejudice to some extent.
So I want to focus in this chapter on how to make Midjourney generate more diverse human bodies.
Method 1: Slider method
In the previous chapter, I introduced the method of adding weights. This method can also help you generate characters of different body shapes. The following is an example from the Midjourney official community:
Plus-sized ginger female dressed in bohemian vest, maxi skirt, espadrilles, walking across autumn campus courtyard, morning atmosphere:: plus-sized::-0.3 --ar 2:3
You can control the parameters after plus-sized to make the characters in the picture fatter or thinner. It should be noted here that negative numbers do not mean thinner, but weaken the plus-sized weight. If your prompt word is thin and you add a negative slider parameter, it means fatter:
Method 2: Direct description
Midjourney knows some diverse words, such as:
- blind
- deaf
- cerebral palsy
- Accessibility: Impaired mobility