3D Image Segmentation-Fully Automatic or Interactive
3D-DOCTOR provides three types of 3D image segmentation functions: Fully Automatic Segmentation, Interactive Segmentation, and Segment Object Using a Training Area.
Need more info on 3D image segmentation and algorithms implemented in 3D-DOCTOR? Please read "Image Segmentation: The First Step In 3-D Imaging" by Dr. Y. Wu.
1. Fully Automatic Segmentation: Open the 3D image and select 3D-DOCTOR's 3D Rendering/Auto Segment function. Enter the number of objects you'd like to extract and that's all you need to do. 3D-DOCTOR uses the 3D image texture to separate objects and trace their boundaries accurately. The boundary data can be edited easily using the vector based Boundary Editor functions and exported to many graphics formats. Most 3D images can be segmented in just seconds or minutes using this method, which will reduce your image processing time significantly. The extracted object boundaries are then used by both the 3D surface rendering and direct volume rendering functions. The picture below shows the segmentation results using the fully automatic segmentation function.
2. Semi-automatic Interactive Segmentation: For each image plane, 3D-DOCTOR first estimates the optimal threshold using the image histogram derived from the current plane. The user can interactively adjust the threshold to make it perfect for the plane and then let 3D-DOCTOR segment the plane and extract all the boundaries. Once the image plane is segmented, simply switch to the next plane, and do the segmentation again. Easy and fun!
3. Segment Object Using a Training Area: This method is best when extracting boundaries from a 3D object which has unique image attributes. It allows a user drawn image region to be used as a training area to segment the 3D image, based on the features generated from the area. This function is often used to extract certain objects inside a larger object, such as bones and blood vessels within the head in an MRI or CT image.