Summary

The website content provides a tutorial on creating AI-generated images that depict dual concepts or opposites within a single image using various prompt wording techniques.

Abstract

The article titled "AI Image How-To — #35" on the undefined website delves into the technique of generating composite images using AI, specifically focusing on creating images that represent opposing ideas or themes. The author, referred to as "MJ and me," demonstrates this through a series of examples, such as a human-angel hybrid and an ancient-futuristic cityscape. The article emphasizes the importance of precise prompt wording, such as using "half with" followed by the opposing elements, to guide the AI in generating the desired split images. It also explores alternative prompt structures, like starting with "half screen" or using "duality between," to achieve varied results. The author notes that while some images were generated quickly, others required multiple attempts to achieve the desired effect, highlighting the iterative nature of the process. The article concludes with an encouragement to experiment with prompts and an invitation to subscribe to the author's Medium page for more insights.

Opinions

The author believes that the complexity of the prompt does not necessarily correlate with the quality of the generated image.
There is a preference for detailed prompts that include the entire instruction rather than partial snippets.
The author expresses a level of perfectionism, mentioning a personal preference for symmetry and the use of "slightly ADD/OCD" to describe their attention to detail in the image generation process.
The AI's ability to interpret prompts and deliver the desired imagery is acknowledged, but the author also recognizes the need for patience and repetition to achieve satisfactory results.
The author suggests that AI is continually learning, implying that user feedback through prompt adjustments contributes to the AI's development and improvement in image generation.
A sense of excitement and encouragement is conveyed as the author invites readers to explore AI's capabilities and create their own awesome images.

AI-GENERATED IMAGES|ART|HOW-TO

AI Image How-To — #35

Dual images! — secrets revealed!

Have you wanted to do opposing pictures within an image? Images of opposites in some fashion?

There are at least a couple of different ways you can word your prompt to do this.

The prompts for these images are not complex. Some images took a while of generating to get a satisfactory image, others appeared rather quickly.

For the image above, I did not put (or test) separations with commas.

This one took longer to generate to not only get an image I liked but to get the division down the center of the face (instead of slightly left of center. Because I’m slightly ADD/OCD.)

I very simply listed half with (insert object) half with (insert opposite).

💥 Yes, it really is that simple.💥

Note: Unlike some others on Medium who show how to create images, I copy and paste the entire prompt — not just give little snippets!

Prompt for the image above:

half with white angel half with human

You get two images for the price of one prompt here. I wanted to show you that MJ will sometimes give a vertical separation and other times a horizontal one.

These were actually the first images I worked on.

The wording for this is just a little bit different from the first image. This one starts with half screen then describes each half, much like the human/angel.

Prompt for the two images above:

half screen half ancient city and half futuristic city

I decided to try using duality between in place of half this and half that.

I did not get, obviously, two images split down the center but more of a comparison of the two.

Using duality, the objects were occasionally listed side by side without a separation.

Prompt for the two images above:

duality between old and new

Here is another example using duality.

This prompt gave amazing results right away.

Prompt for the two images above:

duality between a before and after make-over

I tried the above prompt using half with as I did for the city. AI continued giving make-over images similar to those using duality.

Prompt for the two images above:

half with before make-over half with after make-over

If you don’t get the results you desire, tweak your prompt and try again! AI is continually learning.

Now that you know how — make something AWESOME!

Sweet Chaos — Tickling your senses through fiction writing while shocking you with the outcome! Nothing is as it seems! Ready to read every addicting story leaving you on the edge of your seat? Not a subscriber? Or subscribing without being a referral? Consider having a subscription through me — I get a little boost at no extra charge to you!

Join Medium with my referral link - Sweet Chaos

Read every addicting story from Sweet Chaos to tickle your senses (and thousands more!) Your membership fee directly…

medium.com