US 11,948,677 B2
Hybrid unsupervised and supervised image segmentation model
Soumya Ghose, Niskayuna, NY (US); Jhimli Mitra, Niskayuna, NY (US); Peter M Edic, Albany, NY (US); Prem Venugopal, Clifton Park, NY (US); and Jed Douglas Pack, Glenville, NY (US)
Assigned to GE PRECISION HEALTHCARE LLC, Waukesha, WI (US)
Filed by GE Precision Healthcare LLC, Milwaukee, WI (US)
Filed on Jun. 8, 2021, as Appl. No. 17/342,280.
Prior Publication US 2022/0392616 A1, Dec. 8, 2022
Int. Cl. G06T 7/00 (2017.01); G06N 3/045 (2023.01); G06N 3/088 (2023.01); G06T 7/10 (2017.01); G16H 30/40 (2018.01)
CPC G16H 30/40 (2018.01) [G06N 3/045 (2023.01); G06N 3/088 (2013.01); G06T 7/10 (2017.01); G06T 2207/10081 (2013.01); G06T 2207/20081 (2013.01); G06T 2207/20084 (2013.01); G06T 2207/30101 (2013.01)] 20 Claims
OG exemplary drawing
 
1. A system, comprising:
a processor that executes computer-executable components stored in a computer-readable memory, the computer-executable components comprising:
a receiver component that accesses a computed tomography (CT) image depicting an anatomical structure;
a probability component that generates, via an unsupervised modeling technique, at least one class probability mask of the anatomical structure based on the CT image, wherein each class probability mask of the at least one class probability mask comprises respective probability values for pixels of the CT image indicating respective probabilities of the pixels belonging to an anatomical class associated with a class probability mask of respective anatomical classes associated with the at least one class probability mask; and
an execution component that generates, via a deep-learning model, an image segmentation of the CT image into the respective anatomical classes based on the CT image and the at least one class probability mask.