A DEFORMABLE TEMPLATE MODEL WITH FEATURE TRACKING FOR AUTOMATED IVUS SEGMENTATION
Prakash Manandhar, Chi Hau Chen, Donald O. Thompson, Dale E. Chimenti · AIP conference proceedings · 2010
Intravascular Ultrasound (IVUS) can be used to create a 3D vascular profile of arteries for preventative prediction of Coronary Heart Disease (CHD). Segmentation of individual B‐scan frames is a crucial step for creating profiles. Manual segmentation is too labor intensive to be of routine use. Automated segmentation algorithms are not yet accurate enough. We present a method of tracking features across frames of ultrasound data to increase automated segmentation accuracy using a deformable template model.