Hello everyone and welcome to a very exciting topic!
Today I found out that midjourney already released their image creation transformers to the public.
You can find their homepage and FAQ’s here.
After having a look at the image generation in the public discord channels i tried some input texts to generate a new agent idea for valorant. This is just for entertainment and educational purpose, so i dont own any rights on valorant from Riot Games nor on MidJourney Image Generation.
So after the first very impressing test, i realized the potential for concept artists for this tool.
You can easily create images of an prompt without even do the work yourself.
In order to generate four images you simply write a prompt like the following one.
/imagine new valorant agent samurai stylized
Just wow!
This is insane, by only describing what kind of character we want to create, we receive 4 images to give us a first visual concept of our prompt. After picking the bottom left picture in order to create further variations we receive four new images. We can do this a few times until we receive very polished concept arts like the following two examples.
So after testing out the capabilities of MidJourney we can try to create some different prompts e.g.
/imagine valorant agent samurai Hokusai style, oil paint, black and white
/imagine Valorant Agent blue samurai katana, high detailed, stylized, Riot Games
So what do you think about these visual concept ideas?
What is your opinion on using Image Transformers to create Concept Art?
Let me know in the comments.
Please keep in mind that these pictures have been created on 22th of August 2022.
The Experience described and the results shown are still changing significantly as the MidJourney Team is still working on improvements!