Introduction

Modeling features of your image data can be an important tool to aid in the understanding of the 3 Dimensional nature your data.

Model Categories

3D Models in DeltaVision can be broken up in to two basic categories. Volume models can be used to model things like nuclei, cell boundaries, etc. Line models are suited for looking at metaphase chromosomes or neurons. The approaches for creating these types of models are quite different.

Creating Volume Models

DeltaVision offers many tools for manual and automated extraction of volume model elements. Here are the basic steps in the process:

• Data Preparation: In order for you or an automated algorithm to be able to distinguish features in the data, it may be necessary to prepare the images by filtering, thresholding, or enhancing the contrast. These tools are found primarily in the Filter menu of the main DeltaVision tool bar.
• Identification of Polygons on 2D sections: 2D Polygons are the basic building blocks of Volume models. A facility exists to create and edit polygons (Edit Polygon) which includes both manual and automated tools. If the data thresholds nicely (such that desired features are easily distinguishable when the lower threshold is raised: See Threshold and Find Border) then an automated tool (2D Object Finder) can be used to create polygons. Once a polygon set is created, you can further refine the polygons using Edit Polygon.
  
  An important thing to note as you manipulate polygons is that the polygons "belong" to the tool being used to create them, but are "associated" with a window. For example, if you are using 2D Object Finder to create polygons in Window 1, but quit the 2D Object Finder tool without explicitly saving the polygons as a disk file, they will be lost. However, if a polygon set is active (although perhaps not yet saved), other tools may access them.
  
  Another use for a set of polygons is in the creation of masks (see Cut Mask).
• Associating 2D Polygons as 3D Volume Objects: The next step in the process is to assemble the polygons created on 2D sections as 3D volume objects. The tool used for this task is the Volume Builder. The inputs for the tool are an Image (file or window) and a Polygon Data set. If polygons are actively associated with a window (if, for example, you have just created polygons using 2D Object Finder), the Volume Builder will attach to those polygons when you specify the window as the Input Image. Once the input Image and polygon set are specified, you can choose Create Objects to have Volume Builder attempt to associate 2D Polygons from section to section, thereby creating 3D objects. Tools exist to modify the associations determined by the Volume Builder.
  
  Once the 3D volume objects are created, you can generate several measurements and models from them. Volume Builder can be used to generate tables of 2D and 3D measurements and several types of wireframe and solid model representations of the objects.
• Viewing the Model: If you have used Volume Builder to create wireframe or solid models of the objects, you can then use 3D Model to view and rotate the models in a 3D viewing canvas.

Creating Line Models

Line models can be created that trace objects such as chromosomes or neurons through 3D space. The model, once created, can then be used to extract data from image to perform 3D karyotyping. Line models are created using the following steps:

• Data Preparation: As with the building of Volume models, the extraction of linear models may require some preparation of the image data for you or the automated algorithms to best trace features in 3D. Various tools in DeltaVision's Filter menu may be helpful.
• Tracing Features in 3D: The 3D Model tool is used to manually or automatically trace model objects. To manually add points to the model, you click on features in the Image Window(s). When you start 3D Model, all windows on the screen are "attached" to the 3D Model tool, so you are advised to only have the image that you're interested in represented on the screen. It may be useful to have the standard Z stack loaded in one window and a version of the same image that has been rotated in some way (see Rotate 3D) loaded in another window. Each object is a connected list of points which can be easily manipulated.
  
  Features may be traced in an image using the AutoModel function of 3D Model (found in the Options menu). AutoModel takes a starting point that you specify and searches through the image volume for a path that is constrained by parameters that you can control. You can choose to have AutoModel look for the entire object in one shot, or to do a stepwise search which allows for user intervention to get through areas of noisy data.
• Applying the Model: Aside from the measurements reported by 3D Model during the model creation and editing process, you can use Line models to extract information from the image data. The Straight tool is designed to take Line models and extract tubes of data from the image and display the results in various ways. If applied to chromosomes, this may be a way to do 3D karyotyping.