For backlit portrait panels, convert your photo into a lithophane. For shallow raised plaques, use a relief or emboss. For full three-dimensional shapes, run photogrammetry or single-photo AI reconstruction, then clean up the mesh before printing. Start by picking a high-contrast image for lithophanes and reliefs, or capture a high proportion of overlapping photos for photogrammetry. Expect to repair the mesh afterward, almost no conversion produces a slicer-ready file on the first try.
TL;DR:
- Photogrammetry requires dozens of overlapping photos and extensive post-processing to produce accurate, manifold models, which can be time-consuming.
- Lithophanes and reliefs only need a single high-contrast photo and are low effort, but lithophanes require backlighting to reveal details clearly.
- Single-photo AI reconstruction is quick and low effort but tends to invent unseen geometry, making it less reliable for complex or occluded subjects.
- Mesh repair tools like Blender's 3D Print Toolbox are essential to fix common issues such as holes, flipped normals, and non-manifold edges before printing.
- Automated services can produce accurate, ready-to-print models using AI, ideal for quick turnaround projects or less DIY experience.
Table of Contents
- Why convert a photo, and when each method fits
- Step 1: photo selection and capture best practices
- Step 2: conversion methods for lithophanes, reliefs, photogrammetry, and AI
- Step 3: mesh cleanup and repair to make it watertight
- Step 4: export to STL, confirm scale, and choose slicer settings
- When this workflow succeeds, and where it tends to fail
- How a professional service handles photo to STL conversion
- Rules of thumb for picking the right method
- Order a custom photo-based 3D print from 3D Print Pics
- FAQ
- Sources
Why convert a photo, and when each method fits
A relief turns a flat image into a raised surface, good for plaques, coins, and wall art. A lithophane is a thin, variable-thickness panel that reveals the image when backlit, popular for lamps and night lights. Full 3D reconstruction, through photogrammetry or AI, tries to recreate actual depth and shape, which suits figurines and standalone objects.
Each path trades effort against fidelity:
- Relief: low effort, low cost, moderate fidelity, works from almost any photo.
- Lithophane: low effort, low cost, high fidelity for light-based display, but only readable when lit from behind.
- Photogrammetry: high effort, free or low-cost software, high fidelity when done carefully.
- Single-photo AI reconstruction: low effort, variable cost, inconsistent fidelity depending on the subject.
Matching the method to the goal before you shoot or upload anything saves the most time.
Step 1: photo selection and capture best practices
A modern smartphone camera is sufficient for reliefs, lithophanes, and single-photo AI reconstruction. Photogrammetry benefits from a DSLR or mirrorless camera with manual focus, since consistent sharpness across dozens of frames matters more than raw resolution.
- Shoot in even, diffuse light and avoid harsh shadows that confuse depth estimation.
- Use a plain, uncluttered background, especially for photogrammetry and AI reconstruction.
- Frame the subject fully in the shot and include a ruler or coin for scale reference.
- For lithophanes and reliefs, increase contrast in editing software before conversion, since Formlabs recommends moderate to high contrast images for clearer printed detail.
- For photogrammetry, remove or mask busy backgrounds so the software does not try to reconstruct them.
Pro Tip: Shoot a few extra frames from slightly different angles even for a single-photo workflow. If the first reconstruction misses detail, you will have backup material without a second shoot.
Step 2: conversion methods for lithophanes, reliefs, photogrammetry, and AI
Each method has its own workflow and its own ceiling on realism.
Lithophane. Convert the photo to grayscale, then feed it into a lithophane generator or a manual height-map tool. Keep total thickness between 0.8 and 3.0 millimeters, since Formlabs notes that this range gives the best light transmission and contrast when backlit.
Relief or emboss. Build a height map from the image's contrast, or trace the subject into a vector first for a cleaner edge. Extrude the height map into a 3D relief and add a flat base around 1.5 millimeters thick for structural stability.
Photogrammetry. Capture many overlapping photos around the subject, feed them into a photogrammetry pipeline, and let the software build a point cloud and mesh. The computation itself often runs quickly, but post-processing to make the mesh printable is usually the larger part of the job.
Single-photo AI reconstruction. A single image goes in, and the software guesses the hidden geometry on the back and sides. This works well for frontal subjects like faces or pet portraits, but anything occluded in the original photo will be invented rather than reconstructed.
| Method | Input needed | Typical effort | Best use |
|---|---|---|---|
| Lithophane | 1 photo | Low | Backlit panels, night lights |
| Relief/emboss | 1 photo | Low | Plaques, wall art, coins |
| Photogrammetry | 40 to 80 photos | High | Figurines, objects needing true depth |
| AI single-photo | 1 photo | Low to moderate | Quick figurines, gifts, portraits |
A quick verification checklist for AI reconstructions: check the back of the model for flattened or smeared geometry, confirm proportions look right from a side view, and compare the silhouette against the original photo before committing to a print.
Step 3: mesh cleanup and repair to make it watertight
A watertight mesh has no gaps, holes, or flipped faces. Slicers need this "manifold" geometry to calculate a continuous print path. A generated mesh, especially from photogrammetry or AI reconstruction, rarely arrives watertight.
- Run a quick check in your slicer first. PrusaSlicer and similar tools attempt automatic repairs on import, and many minor issues resolve without extra steps.
- For anything more serious, open the model in Blender and enable the 3D Print Toolbox add-on, which flags non-manifold edges and offers a one-click "Make Manifold" fix.
- Fill remaining holes manually where the toolbox cannot close them automatically.
- Recalculate normals so every face points outward, a common cause of inverted shading and slicer errors.
- Decimate the mesh carefully if the file is too dense, reducing polygon count without collapsing fine detail.
- Retopologize in MeshLab or Meshmixer if the surface is too irregular for a clean decimation.
Severely corrupted meshes sometimes need dedicated repair tools beyond what a slicer offers. Prusa's knowledge base points to Netfabb access through PrusaSlicer on Windows, plus tools like Meshmixer and 3D Builder for cases that automatic repair cannot fix.
Pro Tip: Before committing to a full-size print, run a slicer preview and check for floating shells, odd overhangs, or holes in the bottom layer. A small test print of just the tricky section catches problems a screen preview misses.
Step 4: export to STL, confirm scale, and choose slicer settings
STL remains the standard for 3D printing because it stores geometry only, no textures or color data, which makes it compatible with nearly every slicer in use. Formats like OBJ or GLB keep texture information but usually need conversion to a clean, manifold STL before they will print reliably.
STL files carry no built-in units, so the first move after import is always to check and correct the scale in your slicer. A model generated at the wrong unit assumption can come out ten times too large or too small without warning.
- Orient reliefs and lithophanes flat, back-side down, to avoid unnecessary supports.
- Orient full 3D prints to minimize overhangs, adding supports only where geometry genuinely needs them.
- Prioritize a finer layer height for lithophanes and reliefs, since surface detail depends on it directly.
- Increase wall count or infill on figurines that need structural strength, like handheld gifts.
Reliefs commonly use a thin base plus a modest relief height, while lithophanes print best with a total thickness optimized for light transmission.
Standard PLA handles most photo-derived prints well. Resin or SLA printing captures finer surface detail, which matters most for lithophanes and reliefs where contrast carries the image.
When this workflow succeeds, and where it tends to fail
Photogrammetry struggles with texture-less or reflective surfaces, since the software needs visible detail to match points across photos. Single-photo AI reconstruction struggles with occlusion: anything hidden from the camera, like the back of a head or the underside of an object, gets invented rather than measured.
- Flat, shiny, or uniformly colored subjects often confuse photogrammetry entirely.
- A single occluded angle in an AI reconstruction usually means a guessed, sometimes distorted, back side.
- Photogrammetry takes longer to capture and process but tends to produce more trustworthy geometry.
- AI reconstruction is faster but less predictable, especially on complex subjects.
When a reshoot does not fix the problem, switching to a relief or lithophane sidesteps the depth-reconstruction issue entirely, since neither method tries to guess hidden geometry.
How a professional service handles photo to STL conversion
Some projects are better handed off. 3D Print Pics turns an uploaded photo, drawing, or logo into a 3D model using AI-powered generation, with a preview available before ordering so customers can approve the shape before anything prints. Models come in up to seven color choices in non-toxic PLA, with size options to match the final use, from a desk keepsake to a larger display piece.
This route tends to save the most time on gifts with a deadline, subjects with occluded geometry that trips up DIY reconstruction, or projects needing multiple color variants without manually recoloring a mesh.
Rules of thumb for picking the right method
Backlit and sentimental: lithophane. Flat and decorative: relief. Full shape and willing to shoot dozens of photos: photogrammetry. Fast and forgiving of imperfection: single-photo AI reconstruction. A short checklist worth keeping nearby: confirm your source photo, keep a backup shot, include a scale reference, and always run a small test print before the final one.
— Ian
Order a custom photo-based 3D print from 3D Print Pics
When the DIY pipeline feels like more trial and error than you want, we turn a photo, drawing, or logo into a finished 3D print without the mesh repair, scale checks, or slicer tuning. We generate an AI preview for approval first, then print in non-toxic PLA with customizable size and color options, and offer rush processing when timing matters.
DIY suits anyone who enjoys the process and has time to iterate. For a gift, a pet portrait, or a logo that needs to look right the first time, our custom 3D models start at $49.99. Place your order and approve the preview before we print.
FAQ
What is the easiest way to turn a photo into an STL file?
For most people, a lithophane or relief generator is the fastest path, since both work from a single photo without the overlapping shots photogrammetry requires. Upload your image, adjust contrast, and export, then expect to repair the mesh before slicing.
Why does my converted STL file look broken in the slicer?
Generated meshes from photogrammetry or AI reconstruction often have holes, flipped normals, or non-manifold edges that slicers struggle to interpret. Running Blender's 3D Print Toolbox or a dedicated repair tool before export usually resolves it.
How many photos do I need for photogrammetry?
A typical approach uses many overlapping photos taken around the subject, with sufficient overlap between frames for the software to find matching points. Post-processing the resulting mesh into a printable, watertight model usually takes longer than the capture itself.
Can I just order a 3D print instead of converting the file myself?
Yes, 3D Print Pics converts an uploaded photo into a 3D model with an AI-generated preview, color choices, and non-toxic PLA printing, starting at $49.99 for a custom model. This suits gifts, portraits, or logos where a guaranteed result matters more than doing the conversion yourself.
What thickness should a lithophane be?
A total thickness between 0.8 and 3.0 millimeters gives the best light transmission and image contrast when backlit, according to Formlabs' lithophane guidance. Thinner areas let more light through, which is what reveals the lighter parts of the image.