Blend original art work using generative AI.
Thoughts on Midjourney as a creative tool
Once upon a time I imagined that I could build my own personal neural style transfer (NST) app using some Github notebooks and a 2017 laptop.
I was saved from this foolish folly by a legion of computer programmers and tech companies who have since unleashed the generative artificial intelligence (AI) revolution of design products and tools. The whole world is now in the grip of AI possibilities.
Of the new generative AI tools available for creative types, my favorite is the natural language image generator, Midjourney.
You may have seen the fantastical images created by Midjourney all over the internet and popping-up in stock image repositories, art contests, and photography contests.
My thoughts about the categorization of Midjourney images as art compared to human creations comes down to the story. As I see it, art tells a story of emotion, skill, and intent that gains its value in the eye of the beholder.
A creator can use an image generator like Midjourney as an art generator in the classic sense, but a creator should be transparent about how generative AI is used in their process.
Here is an example of how I used Midjourney’s blend mode to mash up my own artwork.
Blending images using Midjourney
If you are new to Midjourney, their team has instructions on how to set up your Midjourney account.
Once you have established a Midjourney account, here is how you can use their generative AI bot to blend your work.
Access the Midjourney bot in Discord and type /blend
You will be prompted to add images. You can add up to five. For this example, I added the following two images. One was a digital pattern I made, and one was an original drawing.
If you want to change the default settings, type /settings and you will see these options to choose from:
I decided to try MJ version 5 and stick with the base quality. I wasn’t looking for anything too polished rather get some ideas to develop further.
Here are examples of the images generated with my two image uploads. I submitted the two images in different order to see what variations I could get.
The results were a little bit stark and bondage forward, but I decided to make variations and upscale image one from generation 3.
Making the image my own
Many people use Midjourney to create super polished images, spending hours prompting upscaling and re-prompting. For my purpose, I only spent about 10 minutes to get these blends.
My goal was to have the Midjourney bot help me create something new from the combination of my input images but not too far removed from my style.
I do a lot of work using Adobe Illustrator and decided to rework the upscaled image by applying clipping masks and blending modes to incorporate more of the JEFS digital pattern. I also refined the eyes and face.
Here is the result of my efforts:
Final thoughts
What I like about Midjourney’s blend mode is that I can input my own artwork and see where generative AI will take it.
Are the outputs with this method always smooth, gorgeous, photorealistic, steampunk, famous actor, marvel universe, fashionable, futuristic, anime, award-winning AI masterpieces? No. Do I feel the final results I got in this exercise are more me? Yes. I am a part of each reworked pixel.
I hope this example of using Midjourney’s blend mode inspires you to try blending some of your own creations.
If you liked this story, checkout: Generative inspiration: Using generative AI for motivation
Julia Fletcher is founder of JEFS Storytelling Arts, a graphic design studio, where she uses her unique research skills and artistic talents to create custom visual stories that help clients’ increase engagement and promote the education of their audience.