Guides
Image to 3D model in Blender: three methods
Model over a reference, use a displacement relief, or import an AI-generated mesh: three ways to go from an image to a 3D model in Blender.
3 min read

Quick answer
How do you make a 3D model from an image in Blender?
Blender has no automatic image-to-model button. You can model by hand over the image loaded as a reference, use the image as a displacement map to create a relief such as a coin or a lithophane, or generate a mesh from the image with an image-to-3D tool and import it into Blender to refine.
Which method fits your image?
- A photo or render of a real 3D object (a chair, a lamp): Method 1 or 3.
- A flat, high-contrast image you want as a raised surface (a logo plaque, a coin, a lithophane): Method 2.
- Several objects you need quickly: Method 3.
For drawings specifically, 2D drawing to 3D model in Blender covers curve tracing too.
Method 1: model over a reference image
- Front orthographic view (numpad 1). Add › Image › Reference, choose the photo.
- If you have a side photo, add it in Right view (numpad 3).
- Lower the reference's opacity in its Object Data properties so it doesn't distract.
- Add a primitive that matches the main volume, scale it to the photo.
- In Edit Mode, loop cut (Ctrl+R), extrude (E) and move vertices to follow the outline.
- Orbit constantly. Photos have perspective; your model shouldn't copy it blindly.
Pros: complete control, clean topology. Cons: slow. Perspective in photos makes matching tricky; a photo taken from further away with a longer lens is easier to model over.
Method 2: displacement relief
Turns brightness into height. Good for plaques, coins, embossing, terrain from heightmaps and lithophanes.
- Add a Plane. Add a Subdivision Surface modifier (Simple mode) with high levels, or subdivide in Edit Mode many times.
- Add a Displace modifier. Create a new texture, set it to Image, and open your image.
- Adjust Strength for height. Use a high-contrast, greyscale image for predictable results.
- Apply the modifiers when you're happy, and decimate if it's too dense.
Pros: fast, faithful to the image. Cons: only a relief, not a full object. Turn a picture into an STL explains this relief-vs-object distinction for printing.
Method 3: generate a mesh, then refine in Blender
- Upload the photo to an image-to-3D generator. A clean photo matters; see how to take a photo that converts well to 3D.
- Export GLB to keep textures in one file.
- In Blender: File › Import › glTF 2.0.
- Load the original photo as a reference (Method 1, step 1) and compare.
- Fix scale, pivot and any shape errors. How to clean up an AI-generated 3D model has the full routine.
Pros: a full textured model in minutes; Vi3W retopologises it so it's editable. Cons: hidden sides are guesses; you'll check and correct.
Combining methods
A common professional pattern: generate a base (Method 3) for proportions and texture, then remodel the parts that matter by hand (Method 1). Or use a generated mesh as a sculpting base; see image-to-3D as a sculpting base.
Exporting from Blender
Apply transforms (Ctrl+A › All Transforms), then File › Export: glTF for web and Godot, FBX for Unity and Unreal, STL for printing. The Blender integration guide has settings for Vi3W models.
Want Method 3's head start? Upload your image to Vi3W and open the GLB in Blender.
Frequently asked questions
how to make 3d model from photo in blender
Load the photo as a reference image in an orthographic view and model over it, or import a mesh generated from the photo by an image-to-3D tool and refine it. For a flat relief, use the photo as a displacement texture on a subdivided plane.
how to make a 3d model in blender from a picture
Add the picture with Add › Image › Reference, block the shape with primitives that match its outline, refine in Edit Mode, and check from other angles. It gives full control but takes time; generating a base mesh first is faster for complex objects.
Last reviewed

