Natalie May Flamingo
To begin my AI experiments I first gathered a few different images to work with. Two of these images were drawings of my own creation. I wanted to maintain some similarity between the images so that the final produced amalgamation was still legible as a product of its parts. To accomplish this I used the establishing shot from my storyboard project, which is a greyscale drawing of my house from the point of view of standing at the mailbox. I paired this image with the landscape from project 2 which also features a house based loosely on my own, but adds some color and more intricate backgrounds and patterns. With these two images I was able to have a solid base to begin using my own art in conjunction with the AI program.
When these two drawings I had made were plugged into Midjourney using the blend tool, several oddly constructed houses with sprawling hills began to form. My favorite was the image shown here. It shows a modest house on the right which maintains some of the color from my landscape piece. The house is in a fenced yard with a tree. Behind it sit rolling hills and rocky mountains. The colors in this image are more subdued than my landscape, but still maintain the purples and greys prevalent there. The line usage of my own work is maintained well. Overall, this generated image maintained several elements of my own work that I wanted to work with as I melded it with a more heavily AI-generated image.
In order to get this image I followed along with the class tutorial. When changing the original prompt from that of a blug dog eating a hotdog, I decided to change the animal into a flamingo. Flamingoes are usually pink in the wild, so I thought it would be interesting to see a blue one. THe image also reminds me of the blue herons I would so often see growing up. TO generate this image I fed Midjourney a prompt of a blue flamingo eating a hot dog. Then, once I was satisfied with the curve of the neck, the pastels in the background, and the way the hot dog rested in the flamingo's beak I was able to modify the image specifically in the area around the head. I wanted my flamingo to have a hat. Preferably a big floppy sunhat. Of course, the AI decided a much smaller hat would work just fine, and I must agree.
Now that I had my two generated images (one based off my own work and one purely generated by Midjourney) I sought to combine the two. I blended the images together until I found a hat-wearing blue flamingo and house that suited my needs. This image is the result of that blend. In it stands a blue flamingo with a hat that reminds me of an old newsboy. The flamingo stands in front of a house with a colorful sky. The elements I enjoyed most from this generated image was the retention of my own line style and general house shape combined with the fun colors of the original background. The image reminds me of a children's book and I can see a composition like this being used for a story about a postman flamingo delivering mail around a summer town.
Of course, I did not want to just stop at the postman flamingo. While I enjoyed the abstract almost oil painting background, I wanted more influence from my own work, so I reblended it with my landscape image. This produced a composition that much greater combined my own work with the AI's generations. Here my colors and linework were far more prevalent. And while some of the patterns in the hills were lost, I was able to add some flowers to the forground which are very typical of my own work. Overall, it was interesting to see the ways the AI could further push my own ideas and creations to levels and avenues I had not considered before. I am unsure whether I will work with it more in the future, but it was fun to experiment with.