ACQUISITION-ORIENTATION-DEPENDENT FEATURES FOR MODEL-BASED SEGMENTATION OF ULTRASOUND IMAGES

摘要

A model-based segmentation system includes a plurality of clusters (48), each cluster being formed to represent an orientation of a target to be segmented. One or more models (140) are associated with each cluster. The one or more models include an aspect associated with the orientation of the cluster, for example, the appearance of the target to be segmented. A comparison unit (124), configured in memory storage media, is configured to compare an ultra-sound image to the clusters to determine a closest matching orientation and is configured to select the one or more models based upon the cluster with the closest matching orientation. A model adaptation module (126) is configured to adapt the one or more models to the ultrasound image.

主权项

1. An ultrasound imaging system, comprising:
an imaging system configured to be coupled to an ultrasound probe for acquiring ultrasound images of a target to be segmented; a memory operatively associated with the imaging, system, wherein the memory comprises a plurality of clusters of ultrasound images, each cluster comprising a plurality of images associated with an orientation of the target with respect to the ultrasound probe segmented; one or more models associated with each of the plurality of clusters, the one or more models including at least one aspect comprising a plurality of instances, each associated with an orientation of a respective cluster; and wherein the memory further comprises executable instructions for a comparison unit configured to compare an ultrasound image acquired with the ultrasound probe to the clusters to determine a closest matching orientation and further configured to select at least on instance based upon the closest matching orientation; and a model adaptation module configured to adapt the at least one instance to the ultrasound image and identify the orientation of the target to be segmented.