If I want blue bananas – I’ll have them! –
Taking a more in-depth look at ai and text to image generation, I set myself a small challenge. There are one or two keen photographers that I follow on Facebook. They often post beautiful images of their travels. Therefore I wondered if only using their image descriptions I could replicate their imagery using only ai text to image generation. I found the results both intriguing and and frightening in equal measure.
The slider below illustrates this. The images to the left of the slider were described by the photographer who shot them as: ‘Modern Dutch architecture reflecting in water during calm winds’. The image to the right of the slider was an ai generated image from the same description. I think the ai made a pretty good first attempt.
I found this idea quite fascinating. Therefore, I proceeded to try replicating the work of other artists using only text prompts. Below I used the simple prompt: ‘Lighthouse with fireworks’. The ai once again did a pretty good job on a first attempt although with further text refinements it would probably improve further.
Below is yet another example. On the left of the slider is one of my images of trees and on the right is an ai generated image using the text prompt: ‘misty trees’.
I felt again that with only a short text prompt this was another good attempt by ai to recreate ‘misty trees’.
I can no longer see the point of artistic landscape photography, particularly now, when only pushing a few buttons is the requisite. Why make things harder than they need to be? Below are some more of the results yielded from the ‘misty trees’ prompt. I think the ai performed reasonably well.