| CPC G06T 7/0012 (2013.01) [G06V 10/25 (2022.01); G06V 10/44 (2022.01); G06V 10/764 (2022.01); G06V 10/82 (2022.01); G06T 2207/10081 (2013.01); G06T 2207/30096 (2013.01); G06T 2207/30101 (2013.01); G06V 2201/032 (2022.01)] | 25 Claims |

|
1. A computer-implemented method comprising:
receiving one or more input medical images of a vessel of a patient;
defining a lesion in the one or more input medical images;
defining a region of interest around the lesion in the one or more input medical image by:
detecting a centerline of the vessel in the one or more input medical images,
detecting a start marker and an end marker of the lesion along the centerline in the one or more input medical images based on the defining of the lesion, and
defining the region of interest in the one or more input medical images by segmenting a section of the vessel between the start marker and the end marker;
extracting radiomic features from the region of interest, the radiomic features capturing textural patterns of the region of interest;
determining an assessment of the lesion using a machine learning based classifier network based on the radiomic features; and
outputting the assessment of the lesion,
wherein the machine learning based classifier network is constructed by:
extracting one or more radiomic features from a training dataset,
sampling the one or more radiomic features to balance the training dataset,
selecting a feature set from the one or more sampled radiomic features, and
constructing the machine learning based classifier based on the set of features.
|