For those of us not yet acquainted with the tools and process, would you be able to provide some guidance on where to start, if not a "here's how I do it"?
There’s been lots of discussion about this topic here and elsewhere.
The best thread I’ve found (and contributed to) is on the garden railways forum;
A brief summary of the process is;
You need the best photo possible of the figure you require to use as an ‘Image prompt’ then use an AI model generating web site to create a downloadable STL mesh figure you can print (or even get the site to print it for you ?). Also worth noting that in addition to figures AI generation can make easy work of some simple objects too like loads for wagons or street furniture.
You need a good clear photo of the figure in the pose you require preferably against a plain background with no distractions. Older or poor quality photos can now be enhanced easily by AI.
Use Google Gemini to improve and refine photos requires no fee, extra software or photo editing skills, just some understanding of how to write the best instructions to enhance a photo.
Eg
“Please can you refine and increase the detail and resolution of this photo and give the figure a white background.
Please can you refine and increase the detail and resolution of this photo, recreating the missing part of legs and foot
Can you change the angle of view so that we see him square on.
Please can change this figure’s pose to move his arm away from his body
Please can you create a better higher resolution photo realistic image of this man based on this poor quality image ?”
Once you’ve got the best possible image the rest is easy enough.
The photo needs to uploaded to an AI mesh creation web site that creates mesh figures from images (you can use text prompts, but the results are often much poorer) and then it creates the figure for you to download. Some services have a free try out option or an option to use an older version to make some samples to give you an idea about the process, but you won’t be able to download the highest quality results unless you take up a subscription.
To get the ultimate quality you’ll need a subscription.
Meshy.ai is generally regarded as the best currently. There’s usually a half price first month offer, as first time user you’ll also be offered extra free credits for trying features and writing reviews. With some pre-planning you could get hundreds of figures in that first month, so it’s worth getting all the photo prompts ready before starting a subscription. You can also end the sub after a month too.
Tripio3 is almost as good as Meshy and if you sign up to Makerworld you can use it freely.
It makes life easier later if you specify the figure’s height for download. You may also need to run the mesh figure through a repair routine it fix any odd non-manifold errors. MS 3D Builder works well and some slicers have a repair function and will flag up issues before repair.
Then print. Resin is probably the best option, but larger scales G3/16mm/ft etc figures can be successfully printed with a 0.2mm nozzle at 0.05 layer height on a FDM printer with good results.
Above are the basics and the process is really pretty simple to start with.
Beyond that using other 3D software allows further editing; Maybe smoothing creased clothing, some simple edits and, most useful of all, the ability to re-pose figures. Blender is the best known and free, but is a pretty intimidating program, even for experienced computer users. However there are some very good YouTube tutorials that go through the procedures step by step and are reasonably easy to follow.

