Accurate anatomical positioning is essential in magnetic resonance imaging (MRI) to ensure the acquisition of diagnostically useful images. In high-throughput clinical workflows, MRI technicians must position patients rapidly, increasing the likelihood of off-isocenter placements that may necessitate manual repositioning. This study introduces an anatomy-specific, automatic patient positioning control algorithm that predicts required positioning corrections. To support this, we developed and evaluated a segmentation and post-processing pipeline designed to provide actionable feedback to the user. Two annotation strategies – morphological and abstract – were employed. Experimental results show a mean error of 0.35±1.27mm on the shoulder using morphological annotations and 0.50±1.10mm on the wrist using abstract annotations. These results suggest that the proposed approach achieves sufficient accuracy for integration into clinical MRI workflows based on the evaluated datasets.