Photogrammetry, lithophane, or image to 3D? Three ways to get a printable model from a picture

Image to STL — Thu, 05 Nov 2026 15:32:26 +0000

"Can I 3D print this photo?" comes up a lot, and the honest answer is "depends what you mean by print it". There are at least three different things people are asking for, and they need different tools.

I'll be upfront about where I stand: I built imagetostl.io, which does one of the three. But picking the wrong approach wastes more time than picking the wrong tool, so here's how I'd decide.

1. Lithophanes and reliefs: brightness becomes depth

You take one photo, convert brightness into thickness on a flat plate, print it thin in white, and hold it up to a light. Or you make a raised relief for a sign or a coaster.

Good for: photos of people and places, gifts, anything where the picture itself is the point.

Not good for: getting an object. The result is always a plate. Nothing sticks out the back.

Plenty of free generators do this well, and some slicers have it built in. If this is what you want, you don't need anything fancy.

2. Photogrammetry: many photos of a real object

Walk around a physical object, take dozens of overlapping photos, and software reconstructs the geometry.

Good for: copying something that exists and is in front of you. Accurate shapes, real proportions.

Not good for: things you only have one picture of, drawings, characters, or anything that doesn't physically exist. It also struggles with shiny, transparent or plain featureless surfaces, and the cleanup takes time.

3. Image to 3D: one picture, a whole object

This is the newer category. An AI 3D model generator looks at a single image (or a text description) and produces a complete mesh, including the sides and back it never saw.

Good for: turning a sketch, a character drawing, a product photo or a concept into something you can hold. Fast iteration.

Not good for: exact dimensions or mechanical parts. The hidden sides are a reasonable guess, not a measurement. For a bracket that has to fit a bolt, use CAD.

This is the slot Image to STL sits in. You upload a JPG, PNG or WebP (or type a description), it generates the mesh, and you download STL, OBJ or GLB. There's a quick Standard quality for checking the shape (around 80K triangles) and higher tiers, Pro up to about 1.5M triangles and Max, when the shape is right and you want detail. Multi-view input on Pro and Max helps when you do have a few angles. Free tier is 15 credits a month.

Quick decision table

You have	You want	Use
One photo	A glowing picture or a raised plate	Lithophane / relief generator
The real object	An accurate copy	Photogrammetry
One image or just an idea	A full 3D object	Image to 3D
Measurements	A functional part	CAD

Checking the result before you print

Whatever route you pick, look at the model before slicing. For image to 3D output I use the GLB in a tiny three.js page because it's quicker than opening a full editor:

<script type="importmap">
{ "imports": {
  "three": "https://unpkg.com/three@0.160.0/build/three.module.js",
  "three/addons/": "https://unpkg.com/three@0.160.0/examples/jsm/"
} }
</script>
<script type="module">
import * as THREE from 'three';
import { GLTFLoader } from 'three/addons/loaders/GLTFLoader.js';
import { OrbitControls } from 'three/addons/controls/OrbitControls.js';

const renderer = new THREE.WebGLRenderer({ antialias: true });
renderer.setSize(innerWidth, innerHeight);
document.body.appendChild(renderer.domElement);

const scene = new THREE.Scene();
scene.background = new THREE.Color(0x222222);
scene.add(new THREE.HemisphereLight(0xffffff, 0x444444, 2));

const camera = new THREE.PerspectiveCamera(45, innerWidth / innerHeight, 0.01, 100);
camera.position.set(0, 0.5, 2);
new OrbitControls(camera, renderer.domElement);

new GLTFLoader().load('model.glb', (gltf) => {
  const box = new THREE.Box3().setFromObject(gltf.scene);
  gltf.scene.position.sub(box.getCenter(new THREE.Vector3())); // center it
  scene.add(gltf.scene);
});

renderer.setAnimationLoop(() => renderer.render(scene, camera));
</script>

Serve the folder with any static server (npx serve .) because browsers block loading the file directly. Spin it around and look at the back and the underside. That's where generated geometry gets creative.

Then in the slicer: set the real-world scale first (the mesh has no idea how big it should be), check that thin parts are still at least a couple of extrusion widths thick, and plan supports. Organic shapes from any image to STL workflow tend to have lots of overhangs; tree supports are usually kinder to them.

Where I'd draw the line

If your photo is the point, make a lithophane. If the object exists and accuracy matters, scan it. If you only have a picture or an idea and you want a thing you can hold, that's what image to 3D is for, with the understanding that some cleanup may be needed.

I'm still working on how the generator handles busy backgrounds and thin parts, so if you've had a picture turn into something weird, I'd genuinely like to see it.