3D Breast Segmentation for Image Based Shimming

Arjan Willem Simonetti¹, Ronald Holthuizen¹, Chiel Johannes den Harder¹, Mathijs Visser¹, Mark Limkeman¹

This paper proposes an algorithm that performs automatic delineation of the breasts, axillae and implants if present. We aim to perform segmentation on low resolution MR images which are also used for B0 map calculation, and combine the breast segment and B0 map to perform image based shimming. Optimized shimming facilitates the MR workflow by improving the quality of fat suppressed clinical images. This is especially important in the breast which often has a high fat content and a large field of view in bilateral imaging.