Working with Volume Data

Volume data forms the foundation of most Volvicon workflows. This tutorial covers importing data from various sources, applying preprocessing operations, and exporting results for downstream use.

Estimated time: 25 minutes

Prerequisites:

• Completed Getting Started and Navigating the Interface
• Access to volumetric data files (DICOM, NIfTI, or image stacks)

---

Understanding Volume Data

A volume in Volvicon is a three-dimensional grid of voxels (volumetric pixels). Each voxel stores an intensity value representing the measured property at that location—typically X-ray attenuation for CT, signal intensity for MRI, or material density for industrial CT.

Key volume properties include:

Property	Description
Dimensions	Grid size in voxels (X × Y × Z)
Spacing	Physical size of each voxel (mm)
Origin	Position of the first voxel in world coordinates
Data Type	Bit depth and sign (8-bit, 16-bit signed, float, etc.)

---

Importing Volume Data

Volume data can be imported from various sources. The import process automatically extracts or requires specification of spatial metadata (dimensions, spacing, origin).

Importing DICOM Data

DICOM (Digital Imaging and Communications in Medicine) is the standard format for medical imaging:

1. Click File → New → DICOM.
2. Navigate to the folder containing DICOM files.
3. Volvicon scans the directory and lists available series, displaying:
   • Patient name and ID
   • Study and series descriptions
   • Number of slices
   • Modality (CT, MRI, ultrasound, etc.)
4. Select the series to import.
5. Click Load.

Key features:

• Volvicon automatically extracts spatial information (pixel spacing, slice thickness, orientation) from DICOM headers
• Ensures correct physical measurements without manual configuration
• Preserves DICOM metadata for export/compliance

DICOM Orientation

Volvicon attempts to use the correct anatomical orientation from DICOM metadata. However, always verify orientation by checking View → Orientations before proceeding with analysis. Different DICOM sources may encode orientation differently.

Importing 3D Image Files

For structured volume formats with embedded metadata (NIfTI, NRRD, MHD/RAW):

1. Click File → New → 3D Image.
2. Navigate to your file.
3. Select the file and click Open.

Supported formats:

• NIfTI (.nii, .nii.gz) — Neuroimaging standard format
• NRRD (.nrrd) — VTK-compatible format with metadata
• MetaImage (.mhd + .raw) — VTK format in separate files
• NRRD and MHD formats embed spacing and orientation in headers

No additional configuration is required; metadata is automatically extracted.

Importing Image Stacks

Create a 3D volume from a sequence of 2D images:

1. Click File → New → Image Stack.
2. Navigate to the folder containing the image sequence.
3. Select the sequence directory.
4. Volvicon automatically detects subsequent images and presents configuration options.
5. Configure stack parameters:
   • Pixel spacing (X, Y) — Size of each pixel in the image (mm)
   • Slice spacing (Z) — Distance between successive slices (mm)
6. Review the preview (number of slices, final dimensions).
7. Click Load.

Image stack requirements:

• All images must have consistent dimensions (width × height)
• Filenames should sort correctly (use zero-padded numbers: 001, 002, 003 not 1, 2, 3)
• All images must have the same bit depth (8-bit, 16-bit, etc.)
• Recommended: PNG or BMP format (JPEG introduces compression artifacts)

Image Stack Best Practices

Use zero-padded sequential filenames (001.png, 002.png, ..., 100.png) to ensure correct ordering. Single-digit filenames sort incorrectly (1, 10, 2, 20, 3...).

Importing Raw Binary Data

For raw voxel data without headers or standard format:

1. Click File → New → 3D Raw Image.
2. Select your raw binary file.
3. Manually specify all parameters:
   • Dimensions (X, Y, Z) — Number of voxels in each direction
   • Spacing — Physical size of each voxel (mm)
   • Data type — 8-bit unsigned, 16-bit signed, 32-bit float, etc.
   • Byte order — Little-endian or Big-endian (CPU architecture dependent)
   • Header offset — Number of bytes to skip before actual data starts
4. Review the preview.
5. Click Load.

Raw Data Import

This import method requires you to know exact data specifications. Errors in any parameter will produce corrupted results. Refer to the file documentation for correct values.

Volume Properties and Visualization

Viewing Volume Properties

To understand your loaded volume data:

1. Select the volume in the Object Browser.
2. View its properties in the Properties panel:
   • Dimensions — Grid size in voxels (X × Y × Z)
   • Spacing — Physical size of each voxel (mm)
   • Origin — Position of the first voxel in world coordinates
   • Intensity range — Minimum and maximum values in the data
   • Memory usage — Total memory occupied

Adjusting Render Properties

Customize how the volume appears in 2D and 3D views:

1. Select the volume in the Object Browser.
2. Navigate to Image → General → Render Properties.
3. Configure visualization options:
   • Presets — Manage presets for 3D volume rendering.
   • Volume Rendering — Configure blending mode, interpolation, transfer functions, and shading for 3D volume visualization.
   • Masking — Apply masks to the 3D‑rendered volumes.
   • Other Settings — Adjust slice planes, outlines, and clipping options for the 3D view.
4. Changes appear immediately in the views.

---

Geometric Transformations

The Image tab provides tools for modifying volume geometry.

Cropping a Volume

Remove unwanted regions to reduce data size and focus on areas of interest:

1. Navigate to Image → Transform → Crop.
2. The crop tool provides two methods:

Interactive Cropping:

1. Click Enable cropping to enter interactive mode.
2. Drag the crop box handles in 2D or 3D views.
3. The preview shows the resulting region.
4. Click Apply to execute the crop.

Numeric Cropping:

1. Enter exact voxel ranges for X, Y, and Z.
2. Review the resulting dimensions.
3. Click Apply.

Crop Best Practice

Crop early in your workflow to improve performance. Processing smaller volumes is faster and uses less memory.

Resampling a Volume

Change voxel dimensions by up-sampling or down-sampling:

1. Navigate to Image → Transform → Resample.
2. Configure target spacing or dimensions:
   • Pixel spacing — Specify desired voxel size
   • Voxel dimensions — Specify desired grid size
3. Select interpolation method:

Method	Description	Use Case
Nearest	Copies nearest voxel value	Masks, label maps
Linear	Interpolates between neighbors	General volumes
Cubic	Smoother interpolation	High-quality images

4. Click Apply.

Mask Resampling

Always use Nearest interpolation when resampling masks or label maps. Other methods create intermediate values that corrupt discrete labels.

Flipping a Volume

Mirror the volume along an axis:

1. Navigate to Image → Transform → Flip.
2. Select the axis: X, Y, or Z.
3. Select target objects (active, selected, or all volumes).
4. Click Apply.

Padding a Volume

Add voxels around the volume borders:

1. Navigate to Image → Transform → Pad.
2. Specify padding amounts for each face:
   • -X / +X — Left and right padding
   • -Y / +Y — Back and front padding
   • -Z / +Z — Bottom and top padding
3. Click Apply.

Padding is useful when:

• Surface generation requires margin around the mask
• Registration algorithms need border space
• Edge effects occur during filtering

---

Image Processing Operations

Smoothing and Denoising

Reduce noise while preserving important features:

1. Navigate to Image → Operations → Smooth/Denoise.
2. Select a filter:

Filter	Effect	Preserves Edges
Gaussian	Uniform smoothing	No
Median	Removes salt-and-pepper noise	Moderate
Mean	Averages neighborhood	No
Anisotropic Diffusion	Edge-preserving smoothing	Yes
Bilateral	Edge-preserving smoothing	Yes

3. Configure filter parameters (kernel size, iterations).
4. Click Apply.

Denoising Strategy

Start with mild parameters and increase gradually. Over-smoothing can blur important boundaries needed for segmentation.

Morphological Operations

Apply morphological filters to volumes or masks:

1. Navigate to Image → Operations → Morphology Operations.
2. Select the operation:

Operation	Effect
Erode	Shrinks bright regions
Dilate	Expands bright regions
Open	Erode then dilate (removes small bright spots)
Close	Dilate then erode (fills small dark gaps)

3. Set the kernel size.
4. Click Apply.

Combining Volumes

Perform arithmetic operations between two volumes:

1. Navigate to Image → Operations → Combine.
2. Select the operation:
   • Add — Sum intensities
   • Subtract — Difference
   • Multiply — Product
   • Divide — Ratio
   • Mean — Average
   • Min/Max — Minimum or maximum
3. Select the two input volumes.
4. Click Apply.

Intensity Masking

Apply a mask to isolate regions of a volume:

1. Navigate to Image → Operations → Masking.
2. Select the target volume.
3. Select the mask to apply.
4. Configure:
   • Intensity value — Intensity value to assign to voxels outside the mask.
5. Click Apply.

---

Registration (Alignment)

Align multiple datasets to a common coordinate system.

Landmark Registration

Manual alignment using corresponding points:

1. Navigate to Image → Transform → Landmark Registration.
2. Select the Fixed (reference) object.
3. Select the Moving (to be aligned) object(s).
4. Place corresponding landmark points:
   • Click in the fixed object to place a fixed point
   • Click in the moving object to place the corresponding moving point
   • Add at least 3 points for the registration
5. Click Apply to compute and apply the transformation.

Global Registration

Automatic alignment using optimization:

1. Navigate to Image → Transform → Global Registration.
2. Select the fixed and moving objects.
3. Configure registration parameters:
   • Transform mode — Rigid, Affine, or Deformable
4. Click Run to execute the registration.

---

Exporting Volume Data

Export to Standard Formats

1. Click File → Export → Volume Data.
2. Select the volumes to export.
3. Choose the output format:

Format	Extension	Features
MetaImage	.mhd/.raw	VTK-compatible, includes metadata
NIfTI	.nii, .nii.gz	Neuroimaging standard, compressed option
NRRD	.nrrd	Single file, embedded metadata
Raw	.raw	Binary data only

4. Configure format-specific options.
5. Choose the destination and click Export.

Export as Image Stack

Create a folder of 2D images:

1. Click File → Export → Image Stack.
2. Select the volume.
3. Configure:
   • Slice plane — Direction of slice extraction
   • Image — Slice range [From - Until]
   • Output format — PNG, JPG, BMP
4. Choose the destination folder.
5. Click Export.

Export to DICOM

Write volumes back to DICOM format:

1. Click File → Export → DICOM.
2. Select the volume to export.
3. Choose the output directory.
4. Click Export.

DICOM Compliance

Exported DICOM files should be validated before clinical use. Some metadata may differ from the original source.

---

Practical Exercise

Work through this image processing pipeline:

1. Import a DICOM dataset or NIfTI volume.
2. Crop the volume to focus on your region of interest.
3. Apply Gaussian smoothing with a small radius factor (1-2).
4. Adjust Window/Level for optimal visualization.
5. Export the processed volume as NIfTI.
6. Save the project.

---

Best Practices

Data Management

• Keep original data unchanged; work with copies
• Use descriptive names for processed volumes
• Save projects frequently during long sessions
• Document processing parameters for reproducibility

Performance Optimization

• Crop to region of interest early in the workflow
• Use appropriate bit depth (16-bit is usually sufficient)
• Resample large volumes to manageable sizes for exploration
• Close unused projects to free memory

Quality Assurance

• Verify spacing and orientation after import
• Check that transformations produce expected results
• Compare before/after views when applying filters
• Validate exported data in target applications

---

Troubleshooting

Import Issues

Problem	Solution
DICOM files not detected	Ensure files have valid DICOM headers; try DICOM validation tools
Missing slices	Check for incomplete transfers; verify slice count
Incorrect spacing	Manually specify spacing if metadata is missing
Wrong orientation	Use the Reorient tools in Image → Transform

Processing Issues

Problem	Solution
Operation is slow	Work with cropped data; check available memory
Unexpected results	Verify input volume selection; check parameter values
Out of memory	Close other projects; reduce volume size; resample to lower resolution

---

Next Steps

With volume data management mastered, proceed to:

• Threshold Segmentation — Create masks from intensity values
• 3D Visualization Techniques — Advanced rendering options