3D Print from Photo

A photo becomes a printable relief when you turn its brightness into thickness, check the STL in a slicer, and choose settings for your printer.

Image to STL converter

Complete workflow

1. Choose and prepare your photo

Start with a sharp image and a recognizable subject: an evenly lit portrait, simple drawing, or clear silhouette. Crop away distractions. Heightmap mode uses the entire image by default, so its background also affects the surface. For an outline, choose an extrusion mode and check its background settings. Transparency and color tolerance can remove a background, but these controls do not use AI to recognize the subject; matching details inside the subject can also be removed.

Select JPG, PNG, GIF, BMP, WEBP, or HEIC. Animated files use their first frame. The limit is one image up to 20 MiB; resize larger originals first. An image at least 500 pixels wide is recommended. Low, Medium and High detail target 256, 384 and 768 samples along the longer side without enlarging a smaller original. Previews may use fewer samples than exports. Complex geometry has resource limits, and your printer further limits printable detail.

Decide how you want to view the finished piece. A surface relief is read by looking at raised details and the shadows they cast. A lithophane is read with light behind it: thicker areas transmit less light, while thinner areas transmit more. The same thickness map can serve both experiments, but material, orientation, and lighting change the result. The free tool offers Heightmap, Color heightmap, Extrude silhouette, Extrude colors, Cookie cutter and Lithophane. Lithophanes can be flat, arc-shaped, cylindrical or spherical. These modes use the supplied image; none infers the photographed object’s hidden sides or generates a model from a text prompt.

2. Convert the image and inspect the STL

Use the free converter. Decoding, brightness sampling, mesh generation, and export happen in your browser. Your photo is not uploaded for conversion, and you do not need an account or credits. Keep the tab open while it converts. Save a local project if you want to reopen the source image, optional color overlay and settings in another session; projects are not uploaded or saved to an account. Both Color heightmap and Extrude colors accept an optional color overlay. The overlay changes surface colors rather than the source height or region layout; color-region edits still refer to the source image.

In the default Heightmap mode, the starting values are Image width: 100 mm, Base thickness: 0.8 mm, and Total height: 3 mm, with Invert light/dark off. Width controls the horizontal dimension while preserving the photo's aspect ratio. Base thickness is the thinnest backing. Total height sets the upper height limit measured from the underside, including the base; it must not be below the base thickness. With these defaults, the available relief depth is 2.2 mm.

Heightmap mode makes light areas higher by default. Lithophane mode uses its own minimum and maximum thickness and already makes dark areas thicker for backlighting. You do not need to enable Invert for that mapping; Invert reverses it. With the default Linear curve, brightness maps directly to the selected thickness range; other curves and custom depth points change that mapping. The tool does not stretch image contrast automatically. A uniform image produces a flat plate at the height for its brightness and cannot gain details through inversion.

Rotate and zoom the 3D preview. Check the subject, steep transitions, and backing. Check the updated preview after parameter changes, then export the binary STL at the chosen detail level. Standard STL is the default geometry-only format. RGB and BGR options are nonstandard color extensions and are not supported by every slicer. The preview helps you assess shape; check the downloaded file's toolpaths in your slicer too.

3. Check the model in your slicer

Use a slicer compatible with your printer and start from its matching printer and material profiles. Import the STL and check its dimensions before arranging it on the build plate. STL has no unit declaration; the converter writes millimetre coordinates even when you enter dimensions in another supported unit. Confirm the image dimensions and allow for any base, frame or keyring that extends the overall bounds rather than accepting an unexpected import scale.

For an ordinary relief, a flat backing on the bed is a useful first orientation. For a backlit lithophane, standing it upright changes how layer lines cross the image. An upright piece has less contact with the bed and needs an appropriate brim, frame, or other stabilizing approach. Compare both orientations for stability and in the layer preview before committing to a long print; upright is not automatically better.

Slice and move through the layers. Check for a continuous base, toolpaths for small features, and a fit within the build volume. Details may be too small for your nozzle; widening the model or simplifying the photo can help more than increasing relief height.

4. Set layer height, infill, and supports

For an FDM printer with a 0.4 mm nozzle, try a 0.12–0.20 mm layer height when supported by its presets. This is an experimental starting range, not a universal recipe. Finer layers can reduce stepping on slopes but increase printing time; they improve vertical resolution rather than recovering every image detail. See Prusa's layer-height guidance. Resin printers require different profiles.

Choose infill for the purpose. For an ordinary relief, try a suitable profile with 15–20% infill and sufficient solid top and bottom material as an initial test. For a backlit lithophane, 100% infill can help avoid internal shadows; Prusa recommends it for its dedicated lithophane PLA. Check the toolpaths: perimeters may already fill a thin plate, and material changes the result.

Check supports after orientation. A flat backing on the bed often needs none underneath. An upright piece needs stability, but support contact can mark its image. Support overhangs that need it, rather than covering the whole surface. For same-material supports, too small a gap can cause bonding; excessive spacing can leave sagging undersides. Adjust contact distance and separation using a suitable preset and Prusa's support guidance.

Use temperatures, cooling, and speeds from the filament manufacturer's guidance and matching printer preset. Verify them with a small test rather than transferring temperatures from a different material.

5. Print a test, then the finished piece

Print a small version first. Record the photo, thickness range, orientation, material, and slicer settings; change one variable at a time. Reducing width does not reduce thickness automatically, so check every dimension of a miniature test.

Watch the first layer and follow your printer manufacturer's operating guidance. Let the print cool before removal. For a surface relief, inspect it under side lighting. For a lithophane, hold it in front of a cool light source and compare bright and dark regions. If the whole piece is too opaque, test a thinner range or a more translucent material. If it is fragile, add backing thickness rather than relying on the preview alone.

Common problems and fixes

Visible layer lines

Cause: FDM builds a surface in steps, and shallow slopes make those steps easy to see. Orientation and lighting can emphasize them; irregular lines can also reflect printer calibration or extrusion issues.

Fix: Test a finer supported layer-height preset and compare orientation. If lines are uneven rather than regularly spaced, inspect the printer and material profile before changing the relief height.

Supports are difficult to remove

Cause: Support contact is too close, too dense, or placed directly against the image surface. A material profile mismatch can make the support and model bond more strongly than expected.

Fix: Reorient to reduce supported areas, then review interface spacing and contact distance in the appropriate preset. Keep contact away from facial features where possible. Test removal carefully; forcing supports off a thin plate can snap it.

The relief is thin and breaks

Cause: The minimum thickness is too small for the piece's size and handling, or parts of the base do not produce enough continuous toolpaths. A wider plate also flexes more easily.

Fix: Increase Base thickness, keep Total height greater, and regenerate. Examine the backing's toolpaths. Retest light transmission for lithophanes. Lithophane mode offers frame controls for the image edges. Check their dimensions and the attachment in your slicer. A frame can help handling, but it does not replace an appropriate print orientation or support setup.

Printed dimensions are wrong

Cause: Import units, slicer scaling, or printer calibration can change dimensions. With aspect ratio locked, changing width also changes image height. Unlock it to set both dimensions; attachments can extend the overall model bounds.

Fix: Compare slicer dimensions with the requested width in millimetres. Avoid accidental import scaling. If the slicer dimensions are correct but the print differs, check dimensional accuracy with a separate calibration part.

The photo's colors are missing

Cause: Standard STL stores geometry, not the source image’s colors. Color heightmap and Extrude colors show colors in the preview, but RGB/BGR STL color extensions are nonstandard and your slicer may ignore them. In grayscale mapping, different colors with similar brightness can produce similar heights.

Fix: Use Standard STL when geometry compatibility matters. If you try RGB or BGR color STL, first check support in your target software; it does not automatically configure a multicolor print. For a single-material relief, judge the grayscale map and choose filament or paint for the intended appearance.

What kind of photos work best

A front-facing, evenly lit portrait usually gives a clearer first result than a heavily backlit face. Backlighting can make the face a silhouette with little distinction between eyes, nose, and mouth. Choose a photo showing the features you want to preserve.

High contrast helps separate shapes, but fully black shadows or white highlights lose brightness detail. Moderate editing can help; sharpening a blurred photo cannot recover features that were never recorded.

A clean background lets the subject remain the focus. Compare a portrait against a plain wall with the same person in front of dark foliage: the foliage becomes relief too and can compete with the face. Crop first, keep useful contrast, and judge the generated model rather than assuming every attractive photograph will make an equally readable print.

Sample prints

These are planned examples awaiting real photos of physical prints. They are text placeholders, not proof that these pieces have already been printed.

• Portrait lithophane — photo to be added. Show the same finished piece with room lighting and backlighting. Record the image crop, width, thickness range, material, and orientation beside the photos.
• Line-art surface relief — photo to be added. Include a close view under side lighting so readers can inspect fine lines. Record nozzle size and layer height alongside the model settings.
• Thickness comparison — photos to be added. Print the same image with two backing thicknesses, changing no other variable. Show both the handling difference and the effect on transmitted light.

FAQ

Does this reconstruct a full 3D object from one photo?

No. The six free modes build heightmaps, color heightmaps, silhouette extrusions, color-region extrusions, cookie cutters and lithophanes from the supplied pixels and outlines. Curved lithophanes wrap the image onto a surface; they do not recover hidden geometry. Background controls use transparency or color tolerance rather than AI subject recognition. AI 3D generation is planned for the paid tier and is not available yet.

Do I need to register, and where is my photo processed?

Registration is not required for free conversion or STL download. Images and geometry are processed in your browser without uploading the photo for conversion. Save a local project to keep the source image, optional color overlay and settings for another session. These are downloaded files, not cloud storage or account history.

Why does a large photo not produce unlimited detail?

Low, Medium and High detail target 256, 384 and 768 samples on the longer side without enlarging small originals. Preview sampling may be lower than export sampling, and complex geometry has memory and triangle limits. Images are limited to 20 MiB each. Nozzle size, physical dimensions, layer height and material further limit which details survive printing.

Can I use the same settings for a surface relief and a lithophane?

Test the material and settings separately. Heightmap mode makes light areas higher by default and is judged under reflected light. Lithophane mode has its own minimum and maximum thickness, makes dark areas thicker by default and supports flat or curved shapes. Invert reverses that mapping rather than enabling it. Light transmission also depends on the material and print orientation, so print a small test.

Is the downloaded STL ready to send straight to a printer?

Open it in a slicer first. Confirm dimensions and orientation, inspect the layers, and generate instructions using the correct printer and material profiles. Standard STL contains geometry rather than printer instructions. Optional RGB/BGR encodings are nonstandard color extensions and may be ignored by your slicer. Check the exported model and any attached parts; a successful import does not guarantee a successful physical print.

Use the

Image to STL converter

when you are ready to test your photo and compare thickness settings.