Turning a photo of an object into a carve-able 3D model: what's your workflow?

Hi everyone, new member here. I’d like to carve a 3D version of a small object I only have photos of, a carved wooden bird that belonged to my family.

How do you go from a photo of a physical object to something you can generate 3D toolpaths from? Do you build a heightmap from the photo, model it from scratch, or is there a better route? I’m mainly stuck on the step between the photo and a usable model. Any advice on depth and bit choice for that kind of relief would also help.

@ZayleeHenley Welcome to the group.
What CAD/CAM software are you using?

I used EasyCreate for a photo to 3d relief and that worked well. You can adjust the relief depth as well within the web based app. It does cost some money to use it.

@ZayleeHenley Vectric Easy Create. About $20 for 200 credits, which is good for about 50 models.

Look at this post for a few other options
ChatGPT for relief images? - Community Support / Other Software - Sienci Community Forum

The reason that I asked what CAD/CAM application the OP was using is that Easycreate will only export a 3D model in Vectric’s proprietary format V3M. If the OP is not using a Vectric application, Easy Create will also export in a TIFF format, but that is not 3D. It is a bitmap.

I don’t use any Vectric software except for that Easy Create and used it with my Altmill. I’ve only used Fusion with my Altmill currently. I must have exported it as STL.

1 Like

@gwilki Yeah, as @DavidB points out, STL is an option. But it might be new because I don’t recall STL being available last time I used it. It’s there today.

@DavidB @Chucky_ott Good to know. Their website as of this morning only shows V3M as the 3D model format. Here is what I read on their site

“Adjust depth, scale or detail settings of the relief to get the exact result you want and export your model as a standard TIFF file or V3M format to seamlessly import into your VCarve or Aspire.”

I’m glad that they have changed that as I can now use the results on my 3D printer.

Guess they just don’t update their website since I first exported it as STL back in July.

@DavidB Thanks for confirming that.

One thing I would add is to treat the photo as a shape reference, not as a reliable source of scale or depth. If possible, include a known dimension or a ruler in the reference set, and use multiple views so the hidden side is not guessed from lighting alone. For a relief, set the maximum depth deliberately and keep a flat base; brightness in a photo can be surface shading rather than actual height.

After exporting the STL, I would do a pre-CAM check: compare the bounding box with the physical reference, confirm the intended units/scale (STL does not carry unit metadata), check that the base is closed and normals are consistent, and make sure the smallest raised detail is sensible for the ball-nose bit. A quick air-cut or simulation is still useful before cutting the keepsake.

I work on CADProps, so this is a project-related reference rather than an independent recommendation. It is not a photo-to-relief generator, but we keep a small public set of CC0 STEP/STL samples with known dimensions that can be useful for checking an export: CC0 STEP/STL sample files.