Portfolio

Our primary objective here is to showcase and represent a selection of concept designs. These have been empowered by modern/current AI image generation technologies. We have decided to keep this post focused exclusively on static imagery for the moment, however we're exploring emerging video technologies also. Our secondary objective is to probe the limits of the technologies themselves as much as possible around the iterations and experimenting with prompt engineering that is required to refine results.

Approach

We explored several concepts at various depths. Some subjects have been explored at great length. The pursuit of photorealism is paramount in many cases. My goal is to learn how to use the tools available on the market to best represent and assist a design process, from concept through refinement and execution, likely in some other form of media.

As we explored further, several categories have come up. In addition, we have attempted to pursue a bit of a well-rounded approach too, with some of our imagery and categories evolving over time as more samples might be integrated. Also, we wanted to showcase a broad range of skills and potential looks or applications from product, lifestyle, fashion or editorial, science-fiction or futurism, portraiture, technology and more.

In this particular case, we used Playground AI as our primary generation tool. This tool included a few different models built around Stable Diffusion and included several different filters and tweakable parameters that could affect the results. For speed and the ability to generate multiple variations as well as upscale results, we decided upon a paid plan after attempting to gain traction but always having to wait for our free quotas to refresh.

Summary

Please enjoy the selections we've set out for you below. These images represent a small sampling of our current work with generative AI art. Contact us with feedback or to explore new ideas together!

The Playground AI tool has since changed radically into a different thing. We learned they were discontinuing one of the core features, ie, the canvas funtionality on Sept. 30, 2024 and hastened to download the results of our previous year of work or we would lose it all! In having to salvage our efforts, we really got to know the tool, how to use it to it's own benefit, with outpainting, upscaling, erasing, and inpainting, there were lots of new technologies to understand and integrate into a proper workflow.

With the generative side as the primary focus and interface, the first part was only part of the process. Generating new images based on prompts. However, only by trying to make the results usable in some form, did we truly push the technology to its limits and make the results useable, whether by providing proper cropping (ie outpainting), correcting inconsistencies (eraser) or adding new variables or elements (inpainting) to re-generating new runs of imagery with entirely different prompts or other values there were a lot of variables to tweak and new workflows to figure out that were presented in this AI empowered concept art exercise.