Classifying image features of a cardiac images

A method of training a machine learning data processing model for classifying an image feature in order to determine, from cardiac image data, a cardiac characteristic of a patient by evaluating a probability that the image feature matches each class of a predetermined plurality of classes forming a set of classes such as to characterize the image feature, the method of training, further comprising:

• a) receiving, by the machine learning data processing model, a training data set comprising training data, the training data including cardiac image data of at least a part of a cardiac image, and optionally data indicative of a region of interest which includes the image feature;
• b) obtaining, by the machine learning data processing model, a plurality of classes forming a set of classes, wherein the classes together describe status classes of the cardiac characteristic;
• c) receiving, by the machine learning data processing model, a ground truth data set comprising ground truth data indicative of a ground truth class that matches with the image feature;
• d) training the machine learning data processing model based on the training data received in step a), the classes of the set of classes obtained in step b), and the ground truth data received in step c), for enabling the machine learning data processing model, after completion of the training period, to perform the step of providing probability data, the probability data comprising for each class of the set of classes a probability that the image feature matches the respective class.

An electronic image processing system, the system comprising a trained machine learning data processing model, for classifying an image feature in order to determine, from cardiac image data, a cardiac characteristic of a patient by evaluating a probability that the image feature matches each class of a predetermined plurality of classes forming a set of classes such as to characterize the image feature, wherein the system further comprises one or more processors and a memory storing the trained machine learning data processing model and one or more instructions, which when loaded into the memory enable the one or more processors to perform the steps of:

• receiving, by the trained machine learning data processing model, the cardiac image data;
• obtaining, by the trained machine learning data processing model, data indicative of the region of interest;
• collecting, from the trained machine learning data processing model, probability data, the probability data comprising for each class of the set of classes a probability that the image feature matches the respective class;
• evaluating, by a controller, the probability data for the classes of the set of classes, and identifying one or more candidate classes having a significant probability of matching with the image feature;
• determining, by performing a comparison of the probability data for the candidate classes, for each of the identified candidate classes a certainty level indicative of a certainty that the image feature matches the respective candidate class;
• and
• dependent on the step of determining, performing at least one step of:
• when the number of candidate classes equals one, and the certainty level exceeds a threshold level:
• associating the one candidate class with the image feature;
• or
• when the number of candidate classes equals one, and the certainty level is smaller than a threshold level:
• requesting user input indicative of a confirmation or disaffirmation that the one candidate class matches with the image feature;
• or
• when the number of candidate classes exceeds one:
• requesting, for each of the one or more candidate classes, user input indicative of a confirmation or disaffirmation that the respective candidate class matches with the image feature.