Supported Formats

Your data can be uploaded in various medical imaging and 3D formats. This page outlines the most common types we support and what each is used for.

🧠 DICOM (.dcm)

• Standard format used by scanners (CT, MRI, CBCT).
• Includes pixel data and metadata (patient info, modality, acquisition parameters).
• Often stored as a series of 2D slices.

Learn how to upload data on the How to use.

🧊 NIfTI (.nii, .nii.gz)

• Widely used for 3D medical images in AI pipelines.
• Contains voxel spacing, affine orientation, and full volume data.
• Ideal for training, segmentation, and visualization.

🟨 Segmentation Masks (.nii.gz, .nrrd, .png)

Segmentation masks are labeled images that highlight specific regions of interest in a medical scan, such as organs, bones, or tumors.

What is a Mask?

• A mask is a separate image or volume where each pixel/voxel contains an integer label.
• Label values correspond to different structures (e.g., 1 = tooth, 2 = gum).
• Masks can be:
  • Binary: only one structure (values are 0 and 1)
  • Multi-class: multiple regions with different integer values

Format and Use

• Stored in the same format and resolution as the input image (e.g. .nii.gz or .nrrd)
• Used for training AI models, visualizing anatomy, or extracting 3D surfaces (meshes)

Segmentation masks do not contain image intensity — only the labels.

🕸️ Mesh Formats

We support mesh upload and export in the following formats:

STL (.stl)

• Triangle-based 3D surface format.
• Common in surgical planning and 3D printing.
• No color or label support.

PLY (.ply)

• 3D mesh format with support for vertex colors and labels.
• Ideal for AI and colored segmentation output.
• ASCII or binary format accepted.

🧩 Voxel vs. Mesh: What’s the Difference?

Medical imaging data can be represented in two major forms: voxels and meshes.

Voxel Formats (DICOM, NIfTI, Masks)

• Think of voxels as 3D pixels — each one holds a value (intensity or label).
• Voxel data is organized in a grid (width × height × depth).
• Used for segmentation, intensity analysis, and AI training.
• Full internal information is preserved.

Mesh Formats (STL, PLY)

• Meshes represent surfaces, not the full volume.
• Made up of vertices and triangles that form 3D shapes.
• Useful for 3D printing, visualization, and simulation.
• Usually extracted from segmentation masks.

Meshes are lightweight and easier to render. Voxels are richer and needed for AI and clinical interpretation.

✅ Summary

If you’re unsure which format to use, we recommend starting with NIfTI for volumes and STL for meshes.

Need help preparing your files? Contact us