Mesh Generation Module Guide

The Mesh Generation module converts segmented image stacks into 3D surface meshes using the Marching Cubes algorithm. These meshes can be visualized interactively, exported for 3D modeling software, or prepared for 3D printing.

For a conceptual overview, see Mesh Generation.

The Mesh Generation module landing page showing step navigation and file selector

---

Prerequisites

Before using this module, ensure you have:

• A segmented TIFF stack (from Segmentation inference or the Annotation tool)
• The segmentation must contain at least one labeled class (non-background pixels)

---

Quick Start

1. Launch module — Click "Mesh Generation" from the workspace hub
2. Select segmentation — Choose a segmentation result or annotation file
3. Configure — Choose which class(es) to mesh and set the Z voxel scale
4. Generate — Click "Generate Mesh" and wait for processing
5. Use results — Open in 3D Visualization or download the JSON/OBJ files

Other formats: the module writes JSON and OBJ. For STL, PLY, or glTF (GLB), right-click the OBJ in the File Browser and use Convert to…

---

Step-by-Step Guide

Step 1: Data Selection

Select the segmentation data to convert into a 3D mesh.

Step 1 data selection with segmentation file validation

Input Options

Source	Description	When to Use
Recent Results	Segmentation outputs from U-Net inference	After running segmentation inference
Annotation Files	Masks from the Annotation Tool	For manually created segmentations
Uploaded Files	Any labeled TIFF stack	For external segmentation data

File Requirements

• Format: Multi-page TIFF (.tif or .tiff)
• Pixel Values: Integer class labels (0 = background, non-zero = object classes)
• Data Type: Any integer type — uint8, uint16, uint32, int8, int16, or int32. The data is used as-is; no bit-depth conversion is applied

Validation

The module validates your file to ensure it's a proper segmentation:

Check	What It Means
Class Detection	File must contain at least one non-background class
Value Range	Rejects files with too many unique values (likely raw images)
Data Type	Must be integer type, not floating point

Note: Invalid uploaded files are automatically removed from your workspace to keep it clean.

For more details, see Segmentation Data.

Click "Next: Generate Mesh" to proceed to Step 2.

---

Step 2: Mesh Generation

Configure output options and generate your mesh.

Step 2 with data summary, format selection, and class options

Data Summary

Before generating, you'll see a summary of your selected file:

Information	Example
Filename	inference_result.tif
Dimensions	64 × 512 × 512 (slices × height × width)
Detected Classes	Class 1, Class 2
Data Type	uint8
Preview	Thumbnail of middle slice

Output Files

Every run writes two files — no format selection needed:

Format	Description	Use Case
JSON	Slice-based format for the built-in 3D viewer	Interactive visualization in browser (enables the per-class slice-range controls)
OBJ (Wavefront)	Industry-standard 3D geometry	Blender, Maya, 3ds Max, etc.

Need STL, PLY, or glTF? Generate the mesh, then right-click the OBJ file in the File Browser and use Convert to… — STL for 3D printing, PLY, or glTF (GLB).

Class Selection

Choose which classes to include in the mesh:

• All Classes (default) — Generates a separate surface for every detected class
• Single Class — Pick one class value to generate on its own

Z Voxel Scale

Controls the voxel aspect ratio along z relative to the in-plane x/y pixels (range 0.05–20, step 0.1):

Value	Effect
1 (default)	Cubic voxels (1 : 1 : 1)
> 1	Stretches along z — e.g. 2 for a z-step twice the in-plane pixel size
< 1	Compresses along z for finely-sampled stacks

Use this when your acquisition is anisotropic so the mesh keeps true physical proportions. Compute it as z-step / in-plane pixel size from your acquisition metadata. The scale is baked into both the OBJ and the JSON (and carries into STL/PLY/glTF conversions).

For more details, see Z Voxel Scale and Output Options.

Click "Generate Mesh" to start processing.

---

Processing

During mesh generation, you'll see real-time progress updates.

Mesh generation progress with status and elapsed time

Progress Indicators

Element	Description
Status Text	Current operation (e.g., "Processing class 1 of 3")
Progress Bar	Visual percentage complete
Percentage	Numeric progress (0-100%)
Elapsed Time	How long generation has been running

Note: Generation time depends on data size and number of classes. Large volumes may take several minutes.

---

Results

After generation completes, you'll see the results summary.

Results section with mesh statistics and download options

Statistics Displayed

Metric	Description
Total Vertices	Number of 3D points in the mesh
Total Faces	Number of triangular surfaces
Classes Processed	How many classes were meshed
Generation Time	Total processing duration (and the Z voxel scale, when not 1)

Download Options

Download buttons are offered for the JSON and OBJ files.

Actions

• Open in 3D Visualization — Launch the interactive 3D viewer with your mesh
• Start New Run — Return to Step 1 to create another mesh

Tip: If you navigate away while a generation is running, returning to the module resumes it — an in-progress job re-attaches to its progress bar, and a job that finished while you were away shows its results.

---

Output Files

File	Format	Description	Location
Mesh JSON	VoxelSlices	Slice-based mesh for the built-in 3D viewer	results/meshes/
Mesh OBJ	Wavefront	3D geometry for external software	results/meshes/
Metadata	JSON	Generation statistics	results/meshes/

STL, PLY, and glTF (GLB) are produced via the File Browser's Convert to… on the OBJ file.

---

Troubleshooting

Issue	Cause	Solution
"No object classes found"	File contains only background (0 values)	Use a file with actual segmentation labels
"Too many unique values"	File appears to be raw image data	Select a segmentation output, not raw images
Generation takes very long	Large volume or many classes	Normal for large data; wait for completion
Can't click Generate	Z voxel scale out of range	Use a value between 0.05 and 20
File validation failed	Uploaded file is not a valid segmentation	Ensure file is a labeled TIFF with integer values

---

Related Help Articles

Module Overview:

• Mesh Generation — Module introduction

Step-by-Step:

• Segmentation Data — Input requirements and validation
• Output Options — Class selection and outputs
• Z Voxel Scale — Correct proportions for anisotropic stacks