Evaluation of clinical datasets in fully automatic planning assist system for cardiac magnetic resonance imaging

Background Planning assist systems for cardiac MR examinations are necessary for easier operation and shorter examination times. Slice alignment systems [1] have played an important role in achieving these objectives. We propose a new automatic planning assist system for couch adjustment, local shimming, and axial multislice imaging as the input to a slice alignment system [2]. The new system employs an atlas-based segmentation technique using single scout volume data. In the present study, the accuracy and robustness of the proposed method were evaluated based on more than 50 datasets including clinical data, and the results were compared against the degree of interobserver error in manual annotation.