Efficient model-informed co-segmentation of tumors on PET/CT driven by clustering and classification information

Laquan Li, Chuangbo Jiang, Lei Yu, Xianhua Zeng, Shenhai Zheng

Abstract: Highlights•A new 3D model-informed method is designed for tumor co-segmentation in PET/CT images.•An image-based gray histogram region term is proposed.•A new data fidelity term based on Bayesian classification is proposed.•An innovative edge stop function based on FCM algorithm is proposed.