Abstract
To suppress blurring due to cardiac movement, MRI data for the coronary arteries are generally acquired during the diastole rest period. In this research, a physician-specific model of cardiac rest period determination is proposed based on fuzzy classification system. The research aims to mimic and encapsulate the physician-specific decision in our model. From a test dataset, the cardiac rest period was determined based on normalized cross-correlation of consecutive frame images as well as normalized frame number as measured variables. The right coronary artery was used as a region of interest. The modeled physician's decision data determined the fuzzy set thresholds. The thresholds varied between subjects depending on the total number of frames following a linear function derived from the learning dataset. The distance difference of the results between the model and the modeled physician was analyzed, and shown to be significantly closer compared with the distance to other physicians with p<0.05. The results showed that the fuzzy classifier method can be used to model physician decisions.
