Image to STL: AI, heightmap, logo or lithophane?
Choose the right image-to-STL method for a photo, figurine, logo, relief or lithophane, then prepare the mesh for your slicer.
Use AI image-to-STL generation for a full figurine or object, a heightmap for raised surface detail, silhouette extrusion for a flat logo, and a lithophane workflow for a backlit photo. These approaches all produce a 3D mesh, but they interpret the same image differently. Choosing the method is more important than changing a JPG to a PNG.
If you want to hold a miniature of a pet, a relief of its photograph is a different result. If you want a crisp name badge, an inferred sculpted object is usually unnecessary. Start with the physical object you intend to print.
Match the method to your print
| Image and goal | Method to start with | Why it fits | Main check |
|---|---|---|---|
| Pet, toy or character as a full miniature | AI full-3D generation | Estimates a volume around the subject | Hidden sides, legs, ears and connected parts |
| Terrain or an intentionally prepared depth image | Heightmap relief | Maps values to surface height | Height range, base and smoothness |
| Flat logo, sign or keychain | Outline tracing and extrusion | Preserves a deliberate silhouette | Closed paths, small text and holes |
| Portrait viewed with a light behind it | Lithophane | Uses varying thickness to modulate light | Thickness, light source and print orientation |
| Measured replacement component | CAD from measurements | Lets you define exact functional features | Dimensions, tolerances and material |
An STL file is a triangle surface description. The extension does not tell you which method created it, whether the surface is watertight, or whether it will print well at your chosen size.
AI: a full object estimated from a photo
AI image-to-3D generation interprets the subject and produces an estimated shape. It is the relevant route when your goal is a small figure, decorative prop or sculptural object that should make sense from several angles.
For example, a three-quarter photograph of a seated dog can guide the overall body and pose. The back of its ears, the far legs and the underside are still inferred. Inspect those areas before printing. Fur painted into the source photo may not become individual printable strands, and that is often desirable at miniature scale.
In the Image to STL workbench, upload a clear image, sign in, review the generation cost, generate and rotate the model. Set the desired longest side in millimeters and export STL. The image preparation guide explains the supported files and useful viewpoints.
The workbench uses a full-3D geometry workflow. For a raised plaque, traced sign or lithophane, choose a dedicated tool for that method rather than expecting a mode switch in this workbench.
Heightmaps: brightness becomes height
A heightmap treats image values as elevations. A converter maps darker and lighter regions to different heights, according to its settings, and can add a base underneath. This is a direct mapping, not an understanding of which part of a photograph is physically closer.
That distinction matters for ordinary pictures. A white highlight on a dark ceramic bowl may become a raised bump, while a deep shadow may become a depression. The converter is following the brightness data even when the resulting surface does not match the bowl.
Heightmaps work well when you prepare the image as a depth design. For a badge, you could use a flat background value, one value for the raised border and another for the central emblem. Smoothing reduces abrupt steps, but too much smoothing can round off lettering.
ImageToStl describes its online image conversion as a heightmap or logo-to-mesh workflow. It is a separate website from imagetostl.run. Choose the method based on the desired geometry rather than the similarity of the product names.
Logos: trace the outline and give it thickness
For a crisp flat logo, begin with clean artwork. An SVG with closed paths is often a useful source because the outline is already defined. If all you have is a raster PNG, use a high-contrast image and inspect the traced edges.
A practical logo workflow is:
- Simplify the image to the shapes that must print. Remove tiny decorative lines that will disappear at the final size.
- Trace the silhouette or import closed vector paths into a modeling application.
- Extrude those paths to create thickness, and add a base if islands such as letter centers need support.
- Add a hole or attachment feature if the result will be a keychain or tag.
- Export STL, then check dimensions and individual layers in a slicer.
A disconnected dot over a letter or a floating logo element needs a physical connection if the finished object should be one piece. A background that looks transparent in a PNG does not by itself solve that manufacturing problem.
Lithophanes: print for transmitted light
A lithophane uses thickness to control how much light passes through the print. Thicker regions generally appear darker under backlighting; thinner regions transmit more light. The image is intended to be read with light behind it, rather than as a normal shaded sculptural surface.
Use a lithophane-specific converter to map the picture into an appropriate thickness range and shape. Choose those settings for your material, printer and light source. A curved panel, flat frame and night-light insert may need different mounting and stability decisions.
Do a small test when the material or lighting is new. A successful on-screen preview does not establish the contrast of a physical print. Photograph the test under its intended backlight before committing to a larger version.
From generated geometry to a printable STL
Whichever method you choose, export and slicing are separate checks.
- Scale: confirm the imported dimensions. STL does not reliably declare an intrinsic unit.
- Continuity: inspect loose shells, narrow connections and intended openings.
- Thickness: compare the smallest features with your process, material and final scale.
- Supports: consider overhangs and build-plate contact in the chosen orientation.
- Layers: inspect the slicer's preview for missing features, unexpected fills and unsupported islands.
For a textured GLB you want to print, convert a copy to STL. That conversion retains triangle geometry and removes material information. It does not automatically union overlapping objects, close holes or generate supports. Follow the full STL pre-print checklist for the next step.
Common questions
Can any photo become a precise 3D model?
A photo can guide a generated model or a relief, but one view does not provide complete depth measurements. For an accurate replica of a physical object, consider overlapping multi-angle photography, scanning, measurements and manual modeling as appropriate.
Is a heightmap the same as a lithophane?
Both map image values to geometry. A relief is usually read from reflected light and surface shape; a lithophane is designed around transmitted light through varying thickness. Their useful settings and intended presentation differ.
Which method should I use for a portrait?
Use AI generation for a full head or figure, relief for a sculpted plaque, or a lithophane for a backlit photograph. Check the likeness and fine features at the size you intend to print. None of these choices makes the unseen face geometry a measured scan.
Sources and next step
Method references checked November 18, 2026: ImageToStl's conversion explanation and Blender's STL documentation. The method recommendations above are workflow guidance, not a benchmark comparing generated models.
For a full object, start with the AI workbench. For background on how a single image guides a model, read the image-to-3D introduction.