PROBLEM TO BE SOLVED: To provide a mechanism for evaluating quality of automatic image registration.;SOLUTION: A medical registration training component executed in a medical registration system executes a medical registration training operation to a pair of medical studies. When it is determined that the training operation succeeded, the medical registration instance is marked as a positive instance, in medical registration history. When it is determined that the training operation is required to be corrected, the medical registration instance is marked as a negative instance. The medical registration training system trains a failure prediction machine learning model based on the medical registration history using machine learning, so that a failure prediction machine learning model predicts whether or not a new medical registration operation is required to be corrected, and in response to the prediction, performs a step for automatically correcting the new medical registration operation.;SELECTED DRAWING: Figure 3;COPYRIGHT: (C)2019,JPO&INPIT
展开▼
