Transforming UTE-mDixon MR Abdomen-Pelvis Images into CT by Jointly Leveraging Prior Knowledge and Partial Supervision.

Disadvantages of CT include poor soft tissue contrast and exposure to ionizing radiation. While MRI can overcome these disadvantages, it lacks the photon absorption information. Thus, an intelligent transformation from MR to CT is of great interest. To address this need and using combined MR UTE and modified Dixon (mDixon) data, we propose the SCT-PK-PS method that jointly leverages prior knowledge and partial supervision. Two key machine learning techniques: KL-TFCM and LapSVM are used in SCT-PK-PS. The significance of our effort is threefold: 1) Via KL-TFCM, SCT-PK-PS can group the feature data of MR images into five initial clusters of fat, soft tissue, air, bone, and bone marrow. From the initial partitions, clusters needing to be refined are observed and for each of them a few additionally labeled examples are given as the partial supervision for the subsequent LapSVM classification; 2) Exploiting not only the given supervision but also the manifold structure embedded primarily in numerous unlabeled data, LapSVM can obtain multiple desired tissue-recognizers; 3) Jointly using KL-TFCM and LapSVM, and assisted by the edge detector based feature extraction, SCT-PK-PS features good recognition accuracy, which ultimately facilitates the good transformation from MR images to CT images of the abdomen-pelvis.