ROI Pop-Up Menu

Opens the Image Properties panel, in which you can view the basic and advanced properties of the ROI (see Image Properties Panel).

The following options are available to modify and transform regions of interest.

Crop… Opens the Dataset Cropper panel, in which you can crop an object (see Cropping Datasets).

Padding… Opens the Dataset Padding panel, in which you can add padding to an object (see Dataset Padding).

Crop to Bounding Box… Automatically crops the object to the bounding box that encapsulates the labeled pixels of the object.

Sample… Opens the Dataset Sampler panel, in which you can modify spacing by upsampling or downsampling (see Sampling Datasets).

Resample Geometry… Lets you copy the intersected values of the selected object into the geometry, or shape, of another object (see Resampling Geometries). This can be necessary if you need to extract statistics, such as the minimum, maximum, or mean values of corresponding image data (see Statistical Properties), or intend to create a deep learning training set.

Invert… Opens the Dataset Inverter panel, in which you can apply an axis transformation to an ROI or multi-ROI, as well as invert the labeled voxels within an ROI (see Inverting Datasets).

Apply Transformation From… Applies transformations, such as translations, rotations, and scaling, that were applied to a reference object to the selected ROI or multi-ROI (see Applying Transformations from Other Objects).

The following options are available for aligning regions of interest.

With Current View… Automatically rotates the selected ROI so that its orientation in the selected 2D view is aligned to the Z-axis of the world coordinate system.

Axis of Inertia… Lets you automatically align the axis of inertia of the selected region of interest with the axis of inertia of another ROI.

Centroid With… Automatically aligns the centroid of the selected region of interest with the centroid of another object. The centroid of a region of interest is calculated from the bounding box that encapsulates the labeled voxels of the ROI. Applicable objects include images, regions of interest, and multi-ROIs. Reference objects can be selected in the Choose the Object to Align With dialog.

Look at Center of Mass… Automatically aligns the center of mass of the selected region of interest in the current scene view. You should note that the center of mass is calculated from the labeled voxels within the bounding box of the ROI.

Automatically separates single components from an initial segmentation in a manner consistent with the integrity of the connected particles. The output is a multi-ROI with all particles indexed and identified (see Adaptive Watersheds).

Automatically reduces each connected component in a region of interest to a single-voxel wide skeleton, as shown below.

Original ROI (green) and skeletonized ROI (red)

Note For best results, any inner holes in the region of interest should be filled (see Filling Inner Areas).

Note Refer to http://scikit-image.org/docs/dev/api/skimage.morphology.html?highlight=skeleton%203d#skimage.morphology.skeletonize_3d for information about the implementation of this feature.

The following options are available for refining regions of interest.

Process Islands… The Process Islands options — Remove and Isolate by voxel count or rank — let you refine threshold segmentation results or to isolate objects of a certain size (see Processing Islands).

Remove Voxels Below Local Center of Range… Lets you minimize beam hardening effects.

Fill Connected Pores with a Diameter Smaller Than… Lets you fill connected pores with an opening diameter smaller than the selected threshold (see Fill Connected Pores with a Diameter Smaller Than).

Remove Objects That Touch Another ROI or Multi-ROI… Automatically removes objects within the current region of interest that intersect with a selected ROI or multi-ROI (see Remove Objects That Touch Another ROI or Multi-ROI

Create ROI Using Neighbor Count… Creates a new region of interest that contains only labeled voxels with an original neighbor count greater than or equal to a selected threshold (see Create ROI Using Neighbor Count).

The following options are available for creating multi-ROIs of connected components. You can start a connected components analysis after computing measurements for the multi-ROI in which connected voxels are labeled (see Connected Components Analysis).

New Multi-ROI (6-Connected)… Automatically creates a new multi-ROI, in which each group of connected voxels is labeled as a distinct object. Propagation is done by strictly using the 6 faces adjacent to the current seed and will result in the minimum number of connected pixels (see 6-Connected and 26-Connected).

New Multi-ROI (26-Connected)… Automatically creates a new multi-ROI, in which each group of connected voxels is labeled as a distinct object. Propagation is done by strictly using the 6 faces, 12 edges, and 8 corners adjacent to the current seed and will result in the maximum number of connected pixels (see 6-Connected and 26-Connected).

Creates a multi-ROI, in which the labeled voxels of the selected ROI are assigned to a class. In this case, the multi-ROI will contain one class.

Note You can also create a multi-ROI from multiple ROIs (see Creating Multi-ROIs from Regions of Interest).

Creates a multi-ROI, in which the labeled voxels of the selected ROI are assigned to a class and any unlabeled voxels are assigned to an additional class named 'Complement'. In this case, the multi-ROI will contain two classes.

Automatically creates a multi-ROI in which the segments contained within the selected region of interest are labeled as separate connected components. You should note that segments are determined by first skeletonizing the region of interest, applying a watershed, and then expanding the ROI and assigning a label to each branch.

The following options are available for creating distance maps and to map the volume thickness of a region of interest.

Distance Map… Creates a distance map from the selected ROI (see Creating Distance Maps).

Signed Distance Map… Creates a signed distance map from the selected ROI (see Creating Distance Maps).

Geodesic Distance Map… Creates a geodesic distance map using the Fast Marching method in which the shortest path from a seed point that joins points inside a region of interest are mapped. You should note that you need to select a seed point in the Choose a Seed ROI dialog to create a geodesic distance map and that LUTs can be applied to facilitate interpretation, as shown below.

Geodesic distance map of a pore network

Note The geodesic distance is the shortest path fully contained within a region that joins two points, while the Euclidean distance is the straight line distance.

Volume Thickness Map… Automatically computes local thickness for each point in the selected region of interest. Local thickness at is the diameter of the largest sphere that contains and fits into the foreground pixels. If the set of foreground pixels is denoted by , the local thickness is defined by:

Volume thickness maps can be used to determine the diameter distribution of structures in an image. The algorithm follows the same procedure as the Hildebrand-Ruegsegger algorithm, but replaces its sphere plotting phase with a faster algorithm.

The following options are available from generating convex hulls from a region of interest.

A Filled ROI… Generates a convex hull as a filled region of interest from the selected ROI (see Generating Convex Hulls from ROIs).

An Outer Shell ROI… Generates a convex hull as an outlined region of interest from the selected ROI (see Generating Convex Hulls from ROIs).

A Mesh… Generates a convex hull as a mesh from the selected ROI (see Generating Convex Hulls from ROIs).

The following options are available for computing the degree of anisotropy within a region of interest.

Mean Intercept length (MIL)… Lets you compute the degree of anisotropy for the selected region of interest using the mean intercept length (MIL) method (see Computing Anisotropy).

Star Volume Distribution (SVD)… Lets you compute the degree of anisotropy for the selected region of interest using the star volume distribution (SVD) method (see Computing Anisotropy).

Automatically identifies region of interest segments as connected and non-connected porosity (see Connected Porosity).

Note This option is available for Dragonfly 3D World ZEISS edition only. Contact Comet Technologies Canada Inc. for information about the availability of this version of Dragonfly.

The following options are available for generating 3D models from a region of interest.

Generate Contour Mesh… Generates a contour mesh that describes the surface of a region of interest bound to a threshold (see Generating Contour Meshes from ROIs).

Create Sparse Graph… Automatically creates a graph in which the segments contained within the skeletonized region of interest are assigned to edges and vertices. The scalar values of Euclidean Length are computed automatically for each edge (see Creating Sparse and Dense Graphs).

Create Dense Graph… Automatically creates a graph in which the segments contained within the skeletonized region of interest are assigned to edges and vertices. The scalar values of Euclidean Length, Segment Index, Segment Euclidean Length, Segment Length, and Segment Tortuosity are computed automatically (see Creating Sparse and Dense Graphs).

Create an Anisotropy Vector Field… Lets you create a 3D model of anisotropy, in which color scales represent magnitude or direction and density is configurable (see Creating Anisotropy Vector Fields from ROIs and Meshes).

The following options are available for creating new box shapes.

Current Bounding Box… Automatically creates a box that corresponds to the dimensions of the labeled voxels in the selected region of interest.

Current Box… Automatically creates a box that corresponds to the dimensions of the labeled and unlabeled voxels in the selected region of interest.

Current Clip Box… Automatically creates a box that corresponds to the dimensions of the clip box currently applied to the selected region of interest.

In all cases, you can choose to create a box with a specific spacing or granularity in the Create a Box From dialog. For example, in cases in which you will use the box to compute scalar and vector fields.

Create a Box From dialog

Note Whenever you change the number of pixels on the local X, Y, or Z axes, the spacing values will be updated automatically to preserve the original shape of the box. You can also click the Advanced Properties button to open the Advanced Visual Box Properties dialog. You can modify an object's position in space and change its orientation in this dialog (see Advanced Dataset Properties).

Saves the selected region of interest as a template (see ROI Templates).

Lets you create a new ROI that is aligned to the selected viewing angle of the current 2D view (see Deriving New from Current View).

Note In this case, the resulting ROI may have anisotropic spacing.

Lets you create a new multi-ROI with isotropic spacing that is aligned to the selected viewing angle of the current 2D view (see Deriving New from Current View).

The following options are available for exporting regions of interest and the properties of ROIs.

As ORSObject… Exports the selected ROI in the Dragonfly Object file (*.ORSObject extension) format (see Exporting Objects).

Data to File… Exports the basic and statistical properties of the selected ROI in the comma-separated values (*.csv extension) file format (see Basic Properties and Statistical Properties).

To GDT1 Format… Exports the basic and statistical properties of the selected ROI in the GDT1 file format (see Basic Properties and Statistical Properties).

ROI as Binary… Exports the selected region of interest as an 8-bit binary image, in which labeled voxels are assigned a value of 255 and all other voxels 0. Exported data can be saved in the TIFF, JPEG, BMP, or PNG file formats, as well as raw data or in the ORSObject (*.ORSObject extension) file format.

To CZI… Exports ROIs in the CZI file format (*.czi extension), which is a proprietary file format used by ZEISS microscopes to save data. This option is available for Dragonfly 3D World ZEISS edition only. Contact Comet Technologies Canada Inc. for information about the availability of this version of Dragonfly.

Provides a shortcut for selecting macros that can be executed for a single ROI.

Opens the User Data dialog, in which you can view detailed object descriptions, as well as add fields to further describe the selected region of interest (see Managing User Data).

Creates a gridded multi-ROI that includes only voxels that intersect with the selected region of interest (see Making Gridded Multi-ROIs).

Creates a gridded multi-ROI in the same shape as the selected region of interest (see Making Gridded Multi-ROIs).

Creates a new 8-bit binary image from an ROI in which all labeled voxels are assigned a selected value and all unlabeled voxels are assigned a value of 0. You can choose a value for the labeled voxels in the Value for Labeled Voxels dialog

Automatically computes the Euler characteristic number of the selected region of interest.

Lets you automatically update spacing and obtain the correct measurements for images, regions of interest, and multi-ROIs (see Spatial Scale Calibration).

Extracts an object's history as a macro. Extracted macros not only provide an audit trail to help troubleshoot processing issues, but can be edited and replayed to create new objects (see Extracting Object Histories).

The following options are available to modify and transform selected ROIs.

Resample Geometries… Copies the intersected values of the selected ROIs into the geometry (or shape) of another object to create new datasets. Might be necessary if you need to extract statistics from an ROI or multi-ROI, such as the minimum, maximum, or mean values of corresponding image data (see Statistical Properties), or intend to compute other measures that require consistent shapes.

You can choose the object with the required geometry in the Choose Geometry dialog, shown below.

The following options are available to align selected ROIs.

With Current View… Automatically rotates the selected ROIs so that their orientation in the selected 2D view is aligned to the Z-axis of the world coordinate system.

Creates a new multi-ROI, in which the labeled voxels of each selected ROI is assigned to a class. For example, if you create a multi-ROI from five regions of interest, then the multi-ROI will contain five classes (see Creating Multi-ROIs from Regions of Interest).

Creates a new multi-ROI, in which the labeled voxels of each selected ROI is assigned to a class and all unlabeled, or 'complementary' voxels, are assigned to an additional class. For example, if you create a multi-ROI from five regions of interest, and some of the voxels in the regions of interest are not labeled, then the multi-ROI will contain six classes (see Creating Multi-ROIs from Regions of Interest).

The following options are available for creating distance maps from multiple regions of interest.

Create a Distance Map… Creates a distance map from the selected ROIs (see Creating Distance Maps).

Create a Signed Distance Map… Creates a signed distance map from the selected ROIs (see Creating Distance Maps).

Automatically computes the interfacial surface that is common the selected regions of interest. You should note that computations are performed in 3D with interpolated surfaces, not with pixel-wise methods. You should note that interfacial surfaces cannot be computed for overlapping ROIs, that is, ROIs that share labeled voxels.

Note If more than two regions of interest are selected, the interfacial surface will be computed for the first region of interest of interest selected (A) and the union of all other regions of interest selected (B + C + D...).

Note Interfacial surface areas are computed using the Lindblad surface area estimator, as described in Lindblad, J. (2005). Surface area estimation of digitized 3D objects using weighted local configurations. Image and Vision Computing, 23(2), pages 111-122 (DOI:10.1016/j.imavis.2004.06.012).

The following options are available to generate 3D models from selected regions of interest.

Generate Interfacial Mesh… Generates a normal, sampled, or cubic mesh that describes the interfacial surfaces, which can be defined as the point of contact between two phases, for the selected regions of interest. You should note that interfacial surfaces cannot be computed for overlapping ROIs, that is, ROIs that share labeled voxels.

Note If more than two regions of interest are selected, the interfacial surface generated will be between the first ROI selected (A) and the union of all other ROIs selected (B + C + D...).

Note Interfacial mesh representations are approximations only and computed surfaces may not correspond to the actual surface. Computing interfacial surfaces from regions of interest, which are performed in 3D with interpolated surfaces, is the recommended alternative for such an analysis. In some cases, smoothing a mesh may help to align surface measurements to those computed from regions of interest.

Extract Three-Phase Boundary Graph and Compute Throat Thickness… Extracts a graph of the three-phase boundary shared between three selected regions of interest and computes the throat thickness (see Extracting Three-Phase Boundary Graphs).

Extract Three-Phase Boundary Graph… Extracts a graph of the three-phase boundary shared between three selected regions of interest (see Extracting Three-Phase Boundary Graphs).

Saves the selected regions of interest as a template (see ROI Templates).

The following options are available for exporting multiple regions of interest.

As Multiple ORSObjects… Exports each selected region of interest in the Dragonfly Object file (*.ORSObject extension) format to a separate file (see Exporting Objects).

Note If the selected regions of interest share the same name, their file names will be appended with a sequential number.

As ORSObject… Exports the selected ROIs in the Dragonfly Object file (*.ORSObject extension) format (see Exporting Objects).

Data to File… Exports the basic and statistical properties of the selected ROIs in the comma-separated values (*.csv extension) file format (see Basic Properties and Statistical Properties).

To GDT1 Format… Exports the basic and statistical properties of the selected ROIs in the GDT1 file format (see Basic Properties and Statistical Properties).

To CZI… Exports ROIs in the CZI (*.czi extension) file format, which is a proprietary file format used by ZEISS microscopes to save data.

Note This option is available for Dragonfly 3D World ZEISS edition only. Contact Comet Technologies Canada Inc. for information about the availability of this version of Dragonfly.

Provides a shortcut for selecting macros that can be executed for the number of ROIs selected.

Lets you compute a Watershed segmentation, in which the labeled voxels in the selected regions of interest will be used as the seed points and the expanded ROIs will be overwritten into the initial ROIs (see Computing Watersheds).

Note This option is available for Dragonfly 3D World ZEISS edition only. Contact Comet Technologies Canada Inc. for information about the availability of this version of Dragonfly.

Opens the Segmentation Comparator dialog, in which you can compare a region of interest with a ground truth of the same size and shape (see Evaluating Segmentations). For example, to objectively evaluate the prediction or performance of a segmentation model.

Lets you automatically update spacing and obtain the correct measurements for images, regions of interest, and multi-ROIs (see Spatial Scale Calibration).