A.I. Image G-Rated At the Market (1 Viewer)

[Mule]

Here is Meta's offering. I asked for a woman wearing a pink tennis outfit and pink vans at a supermarket. It came up with this. Meta can sometimes come up with photographic quality images, but most of them are cartoonish enough to be spotted immediately as AI.

Meta knows what Vans and Crocs look like. It doesn't know what Converse or Keds look like and, as best as I can tell, there is no way to teach it.

 

[Sim]

Here is Meta's offering. I asked for a woman wearing a pink tennis outfit and pink vans at a supermarket. It came up with this. Meta can sometimes come up with photographic quality images, but most of them are cartoonish enough to be spotted immediately as AI.

Meta knows what Vans and Crocs look like. It doesn't know what Converse or Keds look like and, as best as I can tell, there is no way to teach it.

That’s an interesting take on AI-generated imagery! The pink outfit and Vans do create a vibrant look, even if the overall image leans more towards a playful style rather than hyper-realism. It can definitely be challenging for AI to replicate certain iconic sneaker styles like Converse or Keds, especially given the nuances of branding and design. Have you experimented with other prompts to see how the results vary?

 

[tidybowler]

Here is Meta's offering. I asked for a woman wearing a pink tennis outfit and pink vans at a supermarket. It came up with this. Meta can sometimes come up with photographic quality images, but most of them are cartoonish enough to be spotted immediately as AI.

Meta knows what Vans and Crocs look like. It doesn't know what Converse or Keds look like and, as best as I can tell, there is no way to teach it.

The first thing I noticed with that image is the mismatch between the model's head and body.

 

[iLiKeds]

CivitAI has some really good models that can generate great Vans and very lovely women.

I used to use JIB, but FLUX seems to do a better job most times. Here are samples of both

 

[Mule]

Have you experimented with other prompts to see how the results vary?

Heretofore, most of my work with AI has been text based research. I am pretty good at using it as a tool to assist people who are in a job search and I do use it to write snippets of code to get the syntax I need and then incorporate that into larger programs.

I've not had much luck convincing it to give me good images of any kind.

If I were serious, I would search for tutorials online. I tend to learn by example.