Systems and methods for image segmentation with a multi-stage classifier

A method for identifying a distinctive region in a medical image data set for a patient, comprising:

• applying a first classifier to a medical image data set, the first classifier having a sensitivity configured to identify a first subset of the medical image data set with one or more characteristics similar to a first specified distinctive region, wherein the first classifier uses a Bayesian classification methodology, k-nearest neighbor classification methodology, neural network classification methodology, or any combination thereof;
• passing the identified first subset to a second classifier;
• applying the second classifier to the first subset, the second classifier having a specificity configured to confirm the presence, or absence of the first specified distinctive region within one or more portions of the first subset, wherein the second classifier uses a Bayesian classification methodology, k-nearest neighbor classification methodology, neural network classification methodology, or any combination thereof, and wherein the first specified distinctive region is selected from an occlusive plague region and a vulnerable plague region;
• applying a third classifier to the medical image data set, the third classifier having a sensitivity configured to identify a second subset of the medical image data set with one or more characteristics similar to a second specified distinctive region, wherein the second specified distinctive region is a different type of region than the first specified distinctive region, wherein the third classifier uses a Bayesian classification methodology, k-nearest neighbor classification methodology, neural network classification methodology, or any combination thereof;
• passing the identified second subset to a fourth classifier;
• applying the fourth classifier to the second subset, the fourth classifier having a specificity configured to confirm the presence, or absence, of the second specified distinctive region;
• and
• displaying a medical image based on the medical image data set and showing one or more first specified distinctive regions as identified by the first and second classifiers and one or more second specified distinctive regions as identified by the third and fourth classifiers, wherein the fourth classifier uses a Bayesian classification methodology, k-nearest neighbor classification methodology, neural network classification methodology, or any combination thereof.