Image to 3D reconstructs a 3D model from a reference image — your own artwork or an image you generated in PicoBerry.
Generate from an image
- Switch the workspace to the 3D tool and choose the image input.
- Pick an input mode and upload an image, or select one from your library.
- Set the engine and options (texture, polygon count, license) — the same options as Text to 3D.
- Press Generate.
PicoBerry infers the shape and textures from the image and produces a model in your library.
Input modes
- Single image — reconstruct from one image. Fastest and simplest.
- Multi-view — feed the same subject seen from front, back, left, and right for a more accurate shape. The front view is required; the other angles are optional. (A 4-view sheet from Text to Image makes great input.)
- Batch — turn several images into models one after another. Handy for processing many with the same settings.
Get the best results
- Use a single subject centered on a plain background. Removing the background first makes it cleaner.
- Prefer a clear three-quarter or front view — extreme angles confuse the reconstruction.
- High-contrast, evenly lit images reconstruct more reliably than dark or busy ones.
- For multi-view, keep the subject at the same size and pose across every angle.
After generating
Just like Text to 3D, you can remesh and retexture, animate, and export the result.