Home >> Science & Nature >> Science

How to Segment Images Using Fast Marching

The Fast Marching algorithm is used to quickly create a topology of moving or changing images by representing their boundary conditions as a function of time and speed. The algorithm allows complex 2-D and 3-D images to be generated from an input digital image. It is used in medical imaging applications, such as for generating images of the human brain. Image segmentation is the process of clustering the pixels of a digital image into regions. These regions usually consist of a logical object or part of an object. You can perform Fast Marching Image Segmentation using the "3DSlicer" computer application.

Things You'll Need

3DSlicer
FastMarchingSegmentation module
Digital images
Computer

Instructions

- 1
  Launch the "3DSlicer" application. Since Fast Marching can quickly represent the boundary conditions of irregular images, it is widely used as an image segmentation technique. The "3DSlicer" application has a FastMarchingSegmentation module that allows you to perform fast image segmentation using the Fast Marching algorithm.
- 2
  Click on the ̶0;ID and Initialization Parameters̶1; control panel and input the initialization parameters. You will need to locate and select the image to be segmented and provide a list of points within the image to be segmented as well as the output parameters, including any expected output volume in mL.
- 3
  Click on the ̶0;Run Segmentation̶1; button to start the Fast Marching image segmentation process. Wait until you see an indication that segmentation has completed.
- 4
  Click on the ̶0;Segmentation Result Adjustment̶1; panel and make the required adjustments to the segmented image and parameters. You will see a 3-D rendering of the segmentation in the window. Adjust the parameters until you get your desired segmentation result.
- 5
  Click on the ̶0;Accept Segmentation Result̶1; button to finalize the segmentation once you have finished making all your adjustments and are satisfied that the segmented image meets your criteria. The Fast Marching algorithm uses a manual stopping condition, so you have to visually inspect the results of your segmentation to confirm that it meets your requirements.

Science