Turn one 2D photo into 3D, free and without AI
Drop in one ordinary photo and this tool builds a depth map from visual cues such as height in the frame, focus, haze and the outline of the subject, then turns it into a 3D surface you can tilt, paint and export. It uses classic computer vision instead of AI, and your image never leaves your device.
No sign-up, no upload, no watermark. Exports STL, GLB, PNG and video.
1. How to read depth
Scenes with ground and a horizon: lower in the frame is nearer.
The depth that stays still while the view turns. Set it on your subject.
Opens a gap where the depth jumps instead of stretching the photo, like phone 3D photos.
2. Fix it by hand (optional)
Pick a brush, then paint on the picture. The view turns to face you while a brush is on.
3. Download
How a 2D photo becomes 3D without AI
A camera flattens the world, so a single photo stores colour but no distance. Your brain still sees depth in it, because it reads cues that have nothing to do with stereo vision. This tool reads the same cues, mixes them into a depth map in which white is near and black is far, then raises a mesh of tens of thousands of points by that map and renders it with WebGL.
- 1
Height in the frame
In most scenes the ground recedes towards the horizon, so the lower part of a picture tends to be nearer.
- 2
Focus
Cameras usually focus on the subject. Areas with fine, sharp detail are treated as nearer than soft, blurred ones.
- 3
Aerial perspective
Air scatters light, so distant hills and buildings look paler, bluer and lower in contrast. Hazy areas are pushed back.
- 4
Colour strength
Saturated colours tend to belong to near objects. Faded colours tend to be far away.
- 5
Figure and ground
Colours that touch the border of the image and connect to it are treated as background, so a subject on a plain backdrop separates cleanly.
- 6
Edge snapping
A guided filter aligns the depth map with edges in the photo, so the depth steps happen on outlines instead of smearing across them.
- 7
Shape from silhouette
Round out shapes inflates the subject from its outline with a distance transform, turning a flat cut-out into a rounded form.
Each depth method weights these cues differently. Landscape leans on height in the frame and haze. Subject leans on figure and ground and then rounds the subject out. Light is near and Dark is near skip the cues and use brightness directly, which is what you want for logos, coins, engravings and lithophanes.
What you can make with it
- 3D photos for social posts. Record the video loop for a parallax clip like the 3D photo effect on social apps, from any picture.
- Lithophanes. Print a photo in white filament that shows the picture when light shines through it, for lamps, night lights and gifts.
- Reliefs, signs and stamps. Turn a logo or line drawing into a raised STL for 3D printing or CNC carving.
- Red-cyan anaglyphs. Save a snapshot in Red-cyan 3D view and look at it through cheap paper 3D glasses.
- Assets for 3D scenes. Drop the GLB into Blender, a game engine or a web page as a textured backdrop.
Tips for a better result
- Set the focus plane on your subject, so it stays still while the background moves around it.
- Keep the depth amount low for 3D photos. Around 20 to 30% looks natural; more stretches the edges.
- If the sky comes out near, switch to Landscape or paint it back with Push back.
- For portraits, Subject with Round out shapes at 60% or more gives faces and bodies volume.
- For prints, check the Clay view: it shows the shape without the photo, the way the printer sees it.
Image to 3D questions
Can you really make a 3D image from one 2D photo without AI?
Yes, within limits. A single photo has no measured depth, so any tool has to estimate it. This one uses the monocular depth cues artists and early computer vision relied on: lower in the frame is usually nearer, sharp areas are usually the subject, hazy and pale areas are usually far away, and colours that connect to the border are usually background. The estimate is then snapped to the edges in the photo. It works best on clear subjects and landscapes, and you can paint corrections where it guesses wrong.
Is my photo uploaded anywhere?
No. All processing happens in your browser with JavaScript and WebGL. The image is not sent to FourCents or any other server, which is why it also works offline once the page has loaded.
How do I make a lithophane or 3D print from a photo?
Load the photo, choose the Lithophane view, then use Download STL with lithophane turned on. Thickness follows darkness, so dark areas print thick and block light while light areas print thin. A common starting point is 100 mm wide, 0.8 mm minimum and 3 mm maximum thickness, printed upright at 100% infill in white PLA. For a raised relief or a sign, turn lithophane off and set the relief height and base instead.
What file formats can I export?
STL for 3D printing (a closed, watertight solid sized in millimetres), GLB for 3D viewers, Blender and game engines (the textured surface), a PNG of the depth map, a PNG snapshot of the current view including red-cyan anaglyphs, and a short looping video of the 3D photo for social media.
Which photos give the best 3D result?
Photos with one clear subject against a plain or blurred background, portraits shot in portrait mode, landscapes with a visible foreground and horizon, and high-contrast logos or artwork. Busy, flat-lit scenes and photos taken straight down at a table are the hardest, because they have few depth cues. If your phone saves a depth map, load it with My depth map for the most accurate result.
How is this different from AI depth estimation?
AI depth models learn depth from millions of example images and usually do better on complex scenes, but they are large downloads or need a server. This tool is a few kilobytes of code, runs instantly on a phone, behaves predictably and is fully editable: every cue is a setting you can see, and the depth brush lets you fix anything by hand.
Need 3D, AR or image processing in your own product?
We built this tool in-house. If your website or app needs product viewers, configurators, photo tools or custom WebGL, tell us what you have in mind. The written, fixed-scope quote is free.
Get a free quoteSame tool, other starting points: Lithophane maker · Image to STL · all free tools