Segmentation of ultrasound sequence


Is it possible to segment a structure from a sequence of tracked ultrasound images and create a 3D surface model.

These are the steps I follow:

  1. I stream the optically-tracked ultrasound images from a third-party client using OpenIGTLink.

  2. I then record a Sequence using the tracked ultrasound images.

  3. Using Segment Editor, I create a Segmentation node and add it to the same sequence created above.

  4. Stepping through the sequence, I segment the structure of interest on the ultrasound images. This is turning out to be painful, because I have to change the Source geometry for each ultrasound slice. Is there an easier way to segment the structure on each ultrasound slice.

  5. Once I have segmented the structure on all the ultrasound slices, I would like to combine the segmentations from all the slices in the Sequence to create a 3D model of the structure. How could I do this?


P.S: Volume rendering is not satisfactory to visualize the 3D structure.

Any suggestions to solve this issue.


Maybe this video would be helpful: Tracked ultrasound AI segmentation and 3D reconstruction tutorial (