A Unified Segmentation Method for CT Image Segmentation with Contrast Agent

An important synergy of PET/CT scanners is the use of the CT data for X-ray based attenuation correction of the PET emission data. The attenuation map of PET data can be estimated by partitioning CT image into different tissue types and then assigning the corresponding attenuation coefficients of tissues at 511 keV. The accuracy of attenuation map depends on the tissue classification of CT image. In this paper, we propose a unified segmentation method of CT images with/without contrast agent in PET/CT imaging. The tissue types of CT images can be classified as bone, lung, air, and soft tissue without and with contrast agent. A mixed fuzzy c-mean clustering algorithm is developed to perform automatic segmentation. The performance of the proposed unified segmentation method was evaluated by using clinical CT images. Preliminary experimental results show that the proposed method can successfully separate different tissue types, particularly soft tissue with contrast agent and bone, compared to the segmented results obtained by using the simple intensity-based fuzzy c-means algorithms. We anticipate that the proposed segmentation method is an effective way to obtain CT-based attenuation map for clinical PET/CT quantification.