Manual


Introduction

Welcome to the TurtleSeg manual. This document serves as both a getting started guide as well as a reference. If you're looking for the answer to a particular question, you may find it answered in the FAQ section. Video tutorials are also provided which may make selected concepts clearer.

TurtleSeg is a software tool designed for performing segmentation on 3D images, e.g. medical images such as MRI or CT. The software takes advantage of techniques that allow the user to provide minimal intuitive interaction for guiding the 3D segmentation process. The algorithms are described in the publications section. Segmentation results can be exported as either a 3D surface, or a 3D binary mask image. TurtleSeg is a general-purpose segmentation tool; it is not specialized for segmenting any specific structures. The software is designed to be easy to use and capable of producing accurate segmentations in a small amount of time.

Overview of segmenting 3D images with TurtleSeg

TurtleSeg's main work flow requires the user to contour the object of interest in a sparse set of non-parallel 2D slices. The image data will then be used to combine these "floating" contours together and form a 3D segmentation of the object. If the 3D segmentation is not satisfactory, the user can improve existing contours or add more contours and repeat the process. TurtleSeg assists the user in adding additional contours by providing a mechanism for suggesting to the user a slice plane within which they should contour next to best improve their segmentation. The plane suggestion feature is named Spotlight, in analogy to how theatre spotlights focus the audience's attention to an interesting area on stage.

Document organization

The Getting Started section is organized like a tutorial, gently exposing each feature of TurtleSeg required to segment images. The Window Layout section describes the function of all objects visible in the main TurtleSeg window, including the sidebar options. The Keyboard and Mouse Shortcuts section acts as a reference to the keyboard and mouse functionality, such as camera control and menu shortcuts. The File Formats section gives a quick enumeration of the file formats supported by TurtleSeg.

Getting Started

This section is designed like a tutorial. Most of the topics in this section are also covered by the video walkthrough that can be found in the Videos page.

Starting a new project

TurtleSeg relies on the concept of a Segmentation Project. It is the main workspace within TurtleSeg. For information on Segmentation Projects, see the Managing Projects section. When TurtleSeg is first started, no Segmentation Project is loaded, and so, no work can be done. To begin, a new segmentation project is started by choosing "New Segmentation Project..." from the "File" menu. A Segmentation Project cannot exist without an associated image (the image within which the segmentation will take place), so TurtleSeg now asks the user to choose an image file. You will first be asked to choose between DICOM and non-DICOM formats. If you are loading a non-DICOM format, you will be asked to select the image file. If you choose a DICOM format, you will be asked to select a directory which will be scanned for DICOM images. Once scanned, you can select the DICOM series (3D image) you would like to load from the list. Supported image file formats are listed in the file formats section. It is recommended that the user choose the provided example image of a wrist CT scan named "WristCT.mhd", found in the Example Data folder added to the TurtleSeg Start Menu entry.

Once the image finishes loading, the TurtleSeg window should look something like this:

Adjusting the image intensity mapping

Now that the image is loaded, a typical first step is to adjust the image intensity mapping in order to enhance the contrast between the intensities of the object of interest and the background. Adjust the image intensity mapping by dragging the left and right vertical bars in the Image Intensity Mapping widget (at the top-left of the window) until the object you wish to segment can be seen as clearly as possible. The intensity range can be translated by clicking near the center line and dragging left or right.

An alternative method for adjusting the image intensity mapping is the standard window/level adjustment interface. The user can right click and drag in any of the Slice Views or the 3D View (see Window Layout) in order to adjust the window and level. Dragging left will decrease the window while dragging right will increase it. Dragging down will increase the level while dragging up will decrease it.

Note that the image intensity mapping can be changed at any time during the segmentation process. Also, image intensity mapping changes affect only the currently selected label. For more information on labels, see Multiple segmentations.

For example, the below image shows the result from adjusting the image intensity to facilitate segmenting the wrist radius bone.

Saving the project

Now that the image intensity mapping has been adjusted, and so the segmentation process can begin. Before segmenting anything though, the project should be saved so that the image data does not need to be located again and the image intensity mapping does not need to be set again. This can be done by selecting "Save Segmentation Project" from the "File" menu. Since this is the first time the project is saved, a prompt appears asking for a project file name. Specifying a folder location and file name and then accepting will save the project. The project file will have the extension "tsp", for TurtleSeg Project, which is appended automatically. See the Saving a project section for more information on saving projects. The information saved in the project file essentially allows the user to pick up where he or she left off by loading that file. Note that TurtleSeg has an Autobackup feature in case of power failure.

Creating 2D contours

It's time to start segmenting the object. In the image, find a slice plane that intersects with the object. In order to find a good slice, it may be necessary to scroll the slices along their normal by sliding the scroll bar underneath them, or rotating them by left clicking and dragging any of the three slices in the 3D View. Note: if the three slices are not visible in the 3D View, then their display is toggled off. Try clicking the eye icons in the 3D View (more information). Once a good 2D cross section through your object of interest has been located, start contouring it by double clicking somewhere on the border of the object. Livewire is used to speed up the 2D contouring process, so when the mouse is moved, the contour, or "Livewire", will automatically follow the edge of the object in real time. Left clicking will place a "seedpoint", which has the effect of anchoring the wire to that point. Keep in mind that the view can be panned or zoomed while contouring. Also note that if a mistake is made, it can be undone by hitting CTRL+Z or by selecting "Undo" from the "Edit" menu. Redo is supported through CTRL+SHIFT+Z, or by selecting "Redo" from the "Edit" menu.

As seedpoints are laid along the object's boundary, TurtleSeg learns the appearance of the object you are contouring, thus causing subsequent Livewires to become more accurate.

The contour is completed when it is closed, by clicking near the initial seedpoint. When completed, the contour will change color to indicate that it is done, and it will now also appear in the 3D View. If desired, additional contours can be created at this point as well, but only this first one will be added for this tutorial.

Constructing a 3D segmentation

After at least one plane has been delineated, a 3D segmentation can be created. Select "Build 3D Segmentation" from the "Actions" menu. Upon doing so, a progress bar at the bottom of the window will appear, indicating that TurtleSeg is computing the 3D segmentation. Larger image sizes and slower computers can cause the computation to take longer. The user can cancel the segmentation build if they wish by pressing the ESC key. When complete, the 3D View (bottom-right quadrant in figure) is populated with a 3D surface representing the initial estimate of the 3D segmentation from that single contour. If only a single user contour was provided as the tutorial suggests, a message will be displayed indicating that a poor segmentation was detected. In general, one 2D contour is not enough information for TurtleSeg to properly construct a 3D segmentation, unless the boundaries of the object are perfectly clear and the object's interior is perfectly homogeneous. Another contour is needed, preferably in a plane non-parallel to that of the first contour. While the user is welcome to find a new plane to contour manually, it is recommended that they invoke TurtleSeg's "Spotlight" feature, in order to have it recommend a plane for them to contour.

Building a segmentation affects only the currently selected label.

Spotlight: Automatic contour suggestion

Spotlight is a feature that suggests to the user a plane to segment that improves the resulting 3D segmentation. If the "Poor Segmentation Detected" message is active, Spotlight can be invoked by clicking "Yes". Otherwise, and in general, Spotlight can be invoked by selecting "Suggest Plane" from the "Actions" menu. When invoked, a progress bar indicating that TurtleSeg is calculating the suggested plane will appear. When complete, the Slice Views will be re-oriented, and the bottom left "red" view's border will flash indicating that it shows the suggested slice. Essentially, TurtleSeg is recommending that the user manually contour the bottom left "red" view, similar to how the first contour was created. Note that if the user disagrees with Spotlight's suggestion, they have the choice of choosing their own slice manually, or re-invoking Spotlight to obtain a different suggestion.

Notice the purple line segments that appear now in each Slice View: these are part of the "Turtle Map" for that slice (more technical details are provided in the publications section). In short, the Turtle Map line segments represent the intersection of the slice with the existing contours. While the suggested slice is being contoured, the Turtle Map line segment can be used as a hint to where the object of interest is, and where its borders are, according to the previous user contours.

Segment the suggested slice plane by double clicking somewhere on the radius bone's border, and then trace a contour around the bone. While doing this, you may see the Livewire snap to the Turtle Map end points, causing them to turn green. By placing seedpoints on Turtle Map end points, you are "linking" contours from different planes together, ensuring that they agree with each other. It is not necessary that you link contours together though, TurtleSeg can do this for you automatically if the contours are close enough to each other. When you have finished contouring, select "Build 3D Segmentation" from the "Actions" menu again. If you had not linked all the Turtle Map end points, you will see them turn green as they are linked automatically by TurtleSeg. After the build is complete, the resulting segmentation will look much more like a radius bone.

Editing existing contours

For when inaccurate user contours are noticed, TurtleSeg provides facilities for modifying existing contours. On top of undo/redo support, users can select any existing contour and modify it in any way, or delete it.

To select a contour, click on it in either a Slice View or the 3D View. On selection, the contour will turn blue. Once a contour is selected, double-click on it to orient a slice View with the selected contour, unless a Slice View is already oriented with the contour. The contour should appear within one of the Slice Views, indicated with a flashing border. The seedpoints of the contour should now be visible in the Slice View. Clicking on an existing seedpoint will select it and permit it to be moved. Clicking where a seedpoint does not exist will create a new seedpoint and automatically select it, again allowing it to be moved by dragging with the mouse. As the seedpoints are moved, the Livewire is modified in realtime. Keep adding and/or moving seedpoints until satisfied with the resulting contour. To delete a seedpoint, select it and press the delete key (not the backspace key). To delete the entire contour, press the delete key when the contour is selected but none of its seedpoints are selected. To deselect a seedpoint, click away from the contour. To deselect a contour, click away from the contour when no seedpoints are selected.

Editing seedpoints is one way for the user to link seedpoints to Turtle Map end points. If the user moves a seedpoint near an unlinked Turtle Map end point, the contour will snap to it and turn green. If the user confirms the edit by releasing the left mouse button, the seedpoint becomes linked. If the user attempts to move the seedpoint again, they will notice that it is constrained to move along the Turtle Map line that it is connected to, in order to keep the contours in different planes consistent with each other.

Improving the segmentation further

While the segmentation that has been obtained thus far is taking shape, it is typical that you will need more than two contours to construct a proper 3D segmentation. It can be improved by adding additional contours, and then rebuilding the segmentation. A user's typical workflow is to iterate through the following steps:

  1. Contour the object within a slice (Double-click on the object's border within a slice)
  2. Compute the 3D segmentation (Actions -> Build 3D Segmentation or CTRL+B)
  3. Explore the segmentation in the 3D View and Slice Views. At this point, if the user is satisfied with the results, then they are done.
  4. Choose another plane to segment, either:
    1. manually, by moving the sliders of the Slice Views and/or rotating the slices in the 3D View, or
    2. automatically, by invoking Spotlight to suggest a new slice plane to contour (Actions -> Suggest Plane or CTRL+G)

A typical process for the user to follow, for example, is to repeat the following steps:

  1. Manually contour the current slice
  2. CTRL+B (Build segmentation)
  3. CTRL+G (Invoke Spotlight to suggest a new slice to contour)

Multiple segmentations

TurtleSeg provides a system for segmenting multiple structures within the same image (and project). Since a segmentation can be seen as a labeling of image points (i.e. according to the anatomy they are occupied by), we call separate segmentations "labels". Managing different labels is done through the "Labels" panel in the sidebar (see Labels).

Upon creating a new project, an initial label is automatically created named "New Label". In the labels tab of the sidebar, a drop-down list shows the set of labels including their color, name, and index.

Exporting the segmentation

When the segmentation is complete, it can be exported for further analysis or processing by other software. The export functionality is accessed under the "Export 3D segmentation" submenu of the "File" menu. The user can choose between exporting only the currently selected label's segmentation ("File -> Export 3D segmentation -> Export current label"), or exporting the segmentation for all labels simultaneously ("File -> Export 3D segmentation -> Export all labels...").

Export current label

When exporting the currently selected label, there are two options for the output format of the segmentation:

  • Mesh - Exporting the segmentation as a mesh will save the 3D surface geometry of your segmentation.
  • Binary Mask - Exporting the segmentation as a binary mask will save a new 3D image with the same size and dimensions as the original image. Voxels in the exported image will have the intensity value of either 0 or 1. Voxels with intensity 1 are inside the segmentation, while voxels with intensity 0 are outside.

Export all labels

When the user selects the "Export all labels..." option, a dialogue box will appear offering the following two options:

  • Export all labels to a single file - This option will output all segmentation data to a single label image file. Each pixel of the label image file will have a value equal to the index (e.g. 1,2,3,...) of the label whose segmentation contains that pixel's location. If no segmentation contains a given pixels location, it is assigned the value 0 in the output label image file.
  • Export all labels as separate files - This option automates the process of selecting each separate label one at a time and exporting a binary mesh for it. Multiple files will be generated, so an output folder is required. A base file name will be prepended to each of the generated filenames. The generated filenames will be made unique by appending an underscore and the label index for each different label (e.g. image_01.mhd, image_02.mhd, ...). Clicking OK will show a list of the files that will be created. The export process will begin when the user confirms the output filenames.

See the File Formats section for information on the export file format options.

Managing Projects

A Segmentation Project is the main workspace of TurtleSeg. A Segmentation Project encapsulates the concept of an image and a segmentation within that image.

Creating a new project

A new project can be created by choosing "New Segmentation Project..." from the "File" menu. A segmentation project is associated with an image file by definition, so choosing to start a new segmentation project will result in a prompt for the image which will be associated with the project. If an existing project was already loaded, it will be closed before a new segmentation project is created, since only one segmentation project can be active at a time.

Projects can also be created by choosing an item from the "Recent Images" sub-menu within the "File" menu. This list is populated with the last 6 images the user has opened.

Saving a project

The saved project file will contain the path to the image data, the current intensity mapping, all user contours (including the seedpoints) and the 3D current segmentation. Note that the image data is not part of the project, only a reference to the image's filename is, so the image is not copied. Since the image is referenced and not copied, moving the image file can break the project. Projects can be saved by choosing "Save Segmentation Project" from the file menu, which will prompt the user for a filename to save the project to. The TSP extension is automatically appended to the project filename. Once a filename has been assigned, it will appear in the TurtleSeg title bar and subsequent invocations of "Save Segmentation Project" will save to the same file again. To save to a different filename, choose "Save Segmentation Project As..." from the "File" menu.

Loading an existing project

To load a Segmentation Project, choose "Open Segmentation Project..." from the "File" menu. A prompt will appear asking the user for the TSP file they wish to open. The segmentation project will be opened and the views will be updated with the image associated with that segmentation project. Contours and 3D segmentations should be visible in the loaded project. If a project is already open, opening a new project will close the previous project.

Loading projects can also be done by choosing one of the items from the "Recent Projects" sub-menu of the "File" menu. The list shows the last 6 projects that were opened, where the top item is the most recent.

TurtleSeg TSP files can also be loaded by double clicking on the TSP file in Windows Explorer. The association between TurtleSeg and TSP files is created when TurtleSeg is installed.

Autobackup

A simple autobackup feature is provided. It is active once the user has saved their project for the first time (and so, their project has an associated TSP file). Every time the user computes a 3D segmentation (by choosing "Build 3D Segmentation" from the "Actions" menu, their current project will be saved to the original TSP filename except with ".bak" appended to it. TurtleSeg will open these ".tsp.bak" files as it would a normal TSP file. Autoback up useful in the case of a power failure or forced quit.

Action Menu

When a project is loaded, the "Actions" menu is the access point for much of TurtleSeg's functionality. This section explains what all the action menu items do. The actions menu is shown below.

Build 3D segmentation

As described in the related getting started topic, the "Build 3D Segmentation" action will begin the process of constructing a 3D segmentation from a sparse set of user-approved contours. In order to build a 3D segmentation, at least one user-approved contour must be specified. When invoked, TurtleSeg will first check for, and attempt to automatically correct any unlinked contours (see the Automatically connect contours action). Then the 3D segmentation process will be under way for the currently selected label, during which time the user can adjust the view.

Clear 3D segmentation

The option to clear the 3D segmentation is available after a 3D segmentation has been constructed. It clears the existing 3D segmentation of the currently selected label.

Suggest plane

The suggest plane action becomes available after a 3D segmentation has been constructed. It is detailed in the above Spotlight Getting Started section. Invoking the "Suggest Plane" action immediately after it has been invoked will result in TurtleSeg offering the user with a new, different suggestion. The user can continue to invoke the action to look through many different plane suggestions, in order of decreasing usefulness.

Auto-contour selected plane

The "Auto-contour Selected Plane" option will have TurtleSeg automatically construct contours within the Slice View that the user had most recently interacted with. TurtleSeg will automatically construct a contour by connecting the end points of all Turtle Map line segments. It is possible for the resulting, automatically generated contour, to be incorrect, so the user should double check the results.

Automatically connect contours

The "Automatically Connect Contours" option can be explicitly invoked through the "Actions" menu item. This action is invoked automatically before a 3D segmentation is constructed. All unlinked Turtle Map end points that are near other contours but not intersecting with them will be snapped to the nearby contour, and linked with it. Sometimes TurtleSeg is not able to automatically link contours. If TurtleSeg is unable to link all contours before a 3D segmentation is built, the user is notified and the remaining unlinked contours and given a chance to fix them manually or ignore them.

Image properties

When a project is open, TurtleSeg allows the user to view the properties of the currently loaded 3D image. To do so, choose the "View Image Properties" option of the "File" menu. A popup window opens which displays the location of the image file, as well as the image origin, dimensions and spacing. Finally, the data type associated with each voxel is displayed at the bottom.

Segmentation statistics

When a 3D Segmentation is computed, statistics can be viewed by choosing "File -> View Segmentation Statistics". Currently, segmentation volume is the only statistic available.

Select next/previous generated slice

The user can invoke the actions "Select Next Generated Slice" or "Select Previous Generated Slice" in order to quickly jump from the currently selected generated slice (pertaining to the currently selected label) to the next one. The user can select a generated slice in the first place by clicking on it in the 3D View. If no selected slice is currently selected, the first/last slice will be selected when the user chooses the "Select Next Generated Slice"/"Select Previous Generated Slice" actions, respectively.

Window Layout

Image views

When an image is loaded, the TurtleSeg window will show four different views of it: Three Slice Views and one 3D View.

  1. Slice Views
  2. The Slice Views show 2D slices of the 3D image along 3 different planes. The Slice Views can be translated along their normal by sliding the scroll bar underneath them. The Slice Views are where the 2D contours of the object are created and modified. The image in the view can be flipped horizontally by clicking on the flip button in the bottom right corner of a Slice View. The image can be rotated 90 degrees by clicking the rotate button in the bottom right corner of a Slice View.

  3. 3D View
  4. The 3D View allows the user to see how the 2D slices are oriented with respect to the image's bounding box. The contours added by the user will also be displayed in the 3D View, as well as the segmented surface when it is computed. The user can rotate the 2D slices in this view by left clicking and dragging. There are three colored eye icons in the bottom right corner of the view which toggle whether each of the three planes will be visible in the 3D View or not.

Each view can be resized by clicking and dragging the dividers between the views. By clicking the maximize button in the bottom right of each view, the user can maximize the view.

Sidebar

The sidebar is the menu on the left side of the TurtleSeg window, highlighted here:

The sidebar consists of an image intensity mapping widget at the top, and a set of options underneath it.

Image intensity mapping

See Adjusting the image intensity mapping

Sidebar options

Sidebar options are grouped in to five categories, Reset options, Slice View options, 3D View options, 3D segmentation smoothing and Labels.

Reset options

The buttons in the environment group are for resetting the camera and slice orientations.

  • Reset Windows

    Clicking on the Reset Windows button will reset the window sizes of the main display such that they all occupy an equal amount of area.

  • Reset Cameras

    Clicking on the Reset Cameras button will reset the camera to its initial default position for all three Slice Views as well as the 3D View.

  • Reset Slices

    Clicking the Reset Slices button will reset the slice planes to their initial default positions. Their orientation and translation along their normals will be reset.

  • Reset Intensity Mapping

    Clicking the Reset Intensity Mapping button will re-adjust the intensity mapping to its default setting.

Slice View options

These options affect the display within the Slice Views.

  • Turtle Maps

    The Turtle Map is discussed briefly in the Spotlight section above. It is the intersection of all existing contours with a slice plane, displayed in that slice plane's corresponding Slice View. It can be helpful by giving hints on how to contour an new slice, based on the previous contours.

  • Orientation Cube

    The orientation cube is the cube in the bottom left of each view that indicates to the user the direction the camera for that view is pointed. Checking this option will toggle the visibility of that cube.

  • Crosshairs

    The crosshairs are the indicators showing, for each view, where the other two views are relatively located. Checking this option will toggle the visibility of the crosshairs.

  • Generated Contours

    When a 3D segmentation is built, a series of parallel contours are generated from the pre-existing user contours. The final 3D segmentation surface is constructed based on these generated parallel contours. Checking this option will toggle the visibility of the generated contours in the slice planes, for all labels.

  • Draw Outline and Draw Region

    Toggling the "Outline" option will draw a cyan outline of the currently selected segmentation within each Slice View. The outline is derived from the current 3D segmentation. The toggling the "Region" option will fill the outline with a translucent cyan color.

3D View options

These options affect the display within the 3D View.

  • Image Bounding Box

    Toggle this option to enable the display of the volume image bounding box. When enabled, the bounding box will be displayed by a thin white wireframe.

  • User Contours

    Toggle this option to enable the display of the user specified contours in the 3D View. When enabled, they will be displayed in orange.

  • Orientation Cube

    The orientation cube is the cube in the bottom left of each view that indicates to the user the direction the camera for that view is pointed. Checking this option will toggle the visibility of that cube.

  • Surface

    Toggle this option to enable the display of the 3D Segmentation's surface. It is rendered in cyan with smooth shading. The currently selected label's surface will be completely opaque, while surfaces from other labels will be transparent.

  • Show de-selected label surfaces

    If this option is enabled, surfaces from labels other than the currently selected label will also be rendered, translucently. If it is disabled, only the currently selected label's surface will be shown.

  • Generated Contours

    Toggle this option to enable the display of the generated contours of which the currently selected label's 3D segmentation is derived from. When a 3D segmentation is computed, TurtleSeg automatically generates a dense set of new 2D contours from which the 3D segmentation is derived from. These generated contours can be seen by enabling this option.

3D segmentation smoothing

The scroll bar controls how much smoothing is applied to the mesh, where 0 on the left represents no smoothing, and 9 on the right represents maximum smoothing. Smoothing changes will not take effect until the user chooses to rebuild the 3D segmentation.

Smoothing options affect only the currently selected label.

Labels

The labels menu contains a drop-down list of all labels, as well as buttons to manage the labels. The drop-down list items have the following format:

  • Color - 3D segmentations for this label will have their surface displayed in the 3D View using this color. The color is chosen automatically and cannot be changed.
  • Name - Used to describe and help identify one label from another.
  • Index - The label index is the number in brackets to the right of the label name. It is used when exporting all segmentations to a file.
For managing labels, the user is provided the following options from the sidebar menu:
  • Add Label - By clicking the "Add" button to the right of the drop-down list, the user can create a new, blank label. The new label is assigned the default name, "New Label", and a unique color and index. It is automatically selected when created.
  • Rename Label - By clicking the "Rename" button, an edit box will appear where the user can assign a new name to the currently selected label.
  • Remove Label - By clicking the "Remove" button, the currently selected label will be removed, and all segmentation information associated will be lost (though this action can still be undone).
  • Select Label - By clicking on the drop-down list of labels and clicking on a new label, the user can select a different label. Upon selecting a new label, that label's 3D segmentation (if it exists) will become opaque in the 3D View, user contours associated with that label will become visible, and all subsequent user-contouring and 3D segmentation building will apply only to the newly selected label. Furthermore, the window-level and mesh smoothness options will be adjusted to match the newly selected label-specific settings.

Keyboard and Mouse Shortcuts

Note: A quick keyboard and mouse shortcut reference is available within TurtleSeg, accessible by pressing the F2 key, or selecting the "Keyboard Shortcut Reference" in the "Help" menu.

Actions

Slice View

  • Start Livewiring: Double-click (Left Mouse Button) in any slice plane to begin contouring within that plane, at the click position.
  • Toggle Freehand Mode: While Livewiring, the user can press and hold ALT to enter freehand mode, where the contour will exactly follow the mouse movements.
  • Toggle Polygon Mode: While Livewiring, the user can press and hold CTRL to enter polygon mode, where each subsequent seedpoint will be connected to the previous one by a straight line.
  • Select contour: Single left clicking near a contour will select that contour. Single left clicking near a Turtle Map line segment will select the contour associated with it.

3D View

  • Select contour: Single left clicking near a contour will select that contour.
  • Select generated contour: This action is available when the "Generated Contours" checkbox is enabled in the sidebar 3D View menu. Single left clicking near a generated contour will select that contour.

Both Views

  • Adjust window/level: Right click and drag the mouse in any view to adjust the window/level. See the Adjusting the image intensity mapping section for more information.
  • Delete Contour: Press the delete key when a contour (but not a seedpoint) is currently selected to delete that contour.
  • Delete Seedpoint: Press the delete key when a seedpoint is currently selected to delete that seedpoint. You cannot delete a seedpoint if it would otherwise leave the contour with 2 or less seedpoints.
  • Cancel Segmentation Build: Press the escape key while building a 3D segmentation to cancel the build process.

Camera control

Slice View

  • Adjust View Crosshairs: By single left clicking when the mouse is not near a contour, the other views will be re-positioned so they intersect with the mouse click. The user can also click and hold in order to drag the crosshairs around the view.
  • Pan Camera: Press (Middle Mouse Button) and drag OR Press (SHIFT+Left Mouse Button) and drag.
  • Zoom Camera: Press (CTRL+SHIFT+Left Mouse Button) and drag OR Scroll Mouse Wheel.

3D Window

  • Pan Camera: Press (Middle Mouse Button) and drag OR Press (SHIFT+Left Mouse Button) and drag.
  • Zoom Camera: Press (CTRL+SHIFT+Left Mouse Button) and drag OR Scroll Mouse Wheel.
  • Rotate Camera: Press (Left Mouse Button) and drag.

Adjusting slice planes

  • Translate Plane along its Normal: Press (Left Mouse Button) on the appropriate Slice View's bottom scroll bar and drag left or right.
  • Rotate Planes: Press (CTRL+Left Mouse Button) and drag in the 3D window. Rotation follows the arcball principle (more information).

Color Legend

In TurtleSeg, objects such as contours and seedpoints are colored differently to indicate the different types or states they are in. We provide here a color legend to explain what the different colors mean.

Contour color legend

Color Meaning
Green
The active part of a contour currently being created. It will follow the mouse cursor.
Red
Part of a contour being created that is frozen because a seedpoint was placed.
Orange
A completed, unselected user contour.
Blue
A completed user contour that is currently selected.
Pink
An automatically generated contour.

Seedpoint color legend

Color Meaning
Green
A non-selected seedpoint.
Blue
The highlight around a selected seedpoint.

Other objects

Color Meaning
Dark Purple
A Turtle Map line segment on a slice plane that contains no contours.
Purple
An unlinked Turtle Map line segment.
Green
A linked Turtle Map line segment.

Supported Image File Formats

Supported Input Image Formats

Supported Mesh Export Formats

Supported Binary Mask Export Formats