Multiple operating point false positive removal for lesion identification

A computer program product comprising a computer readable storage medium having a computer readable program stored therein, wherein the computer readable program, when executed on a computing device, causes the computing device to implement a false positive removal engine for removing falsely identified objects in a listing of objects input data set, and to:

• receive, by the false positive removal engine, a received input comprising a listing of detected objects in one or more images of an input volume of images;
• configure, by the false positive removal engine, at least one machine learning classifier computer model, by executing a first machine learning training operation, using a compilation of image volume data, to classify objects as to whether they are true positive object detections, based on first operational parameters to implement a first operating point for the at least one machine learning classifier computer model, and thereby generate a first operating point configured model;
• configure, by the false positive removal engine, the at least one machine learning classifier computer model, by executing a second machine learning training operation, using the compilation of image volume data, based on second operational parameters to implement a second operating point, different from the first operating point, for the at least one machine learning classifier computer model, and thereby generate a second operating point configured model;
• process, by the first operating point configured model, the received input to classify each detected object in the listing of objects as to whether or not the detected object is a true positive or a false positive to thereby generate a first set of detected objects classified as true positive;
• determine, by logic of the false positive removal engine, whether the first set is empty;
• in response to the first set being empty, outputting, by the false positive removal engine, the first set as a filtered list of objects;
• and
• in response to the first set not being empty, process, by the second operating point configured model, the received input to classify each detected object in the listing of objects as to whether or not the detected object is a true positive or a false positive to thereby generate a second set of detected objects classified as true positive and outputting, by the false positive removal engine, the second set of detected objects as the filtered list of objects.

An apparatus comprising:

• a processor;
• and
• a memory coupled to the processor, wherein the memory comprises instructions which, when executed by the processor, cause the processor to implement a false positive removal engine for removing falsely identified objects in a listing of objects input data set, and to:
• receive, by the false positive removal engine, a received input comprising a listing of detected objects in one or more images of an input volume of images;
• configure, by the false positive removal engine, at least one machine learning classifier computer model, by executing a first machine learning training operation, using a compilation of image volume data, to classify objects as to whether they are true positive object detections, based on first operational parameters to implement a first operating point for the at least one machine learning classifier computer model, and thereby generate a first operating point configured model;
• configure, by the false positive removal engine, the at least one machine learning classifier computer model, by executing a second machine learning training operation, using the compilation of image volume data, based on second operational parameters to implement a second operating point, different from the first operating point, for the at least one machine learning classifier computer model, and thereby generate a second operating point configured model;
• process, by the first operating point configured model, the received input to classify each detected object in the listing of objects as to whether or not the detected object is a true positive or a false positive to thereby generate a first set of detected objects classified as true positive;
• determine, by logic of the false positive removal engine, whether the first set is empty;
• in response to the first set being empty, outputting, by the false positive removal engine, the first set as a filtered list of objects;
• and
• in response to the first set not being empty, process, by the second operating point configured model, the received input to classify each detected object in the listing of objects as to whether or not the detected object is a true positive or a false positive to thereby generate a second set of detected objects classified as true positive and outputting, by the false positive removal engine, the second set of detected objects as the filtered list of objects.