3D Colon Object Reconstruction Framework in CT Colonography

Colorectal cancer, also called colon cancer or bowel cancer, includes cancerous growths in the colon, rectum and appendix. It is the second leading cause of death caused by cancers and the third most common form of cancer in the United States [1]. Computer tomographic colonography (CTC) is a minimally invasive technique and rapidly evolving diagnostic tool for the location, detection and identification of benign polyps in the early stage before their malignant transformation. Computer aided detection (CAD) for colonic polyp is a rapidly evolving diagnostic tool for the location, detection and identification of benign polyps on the colon wall on the early stage before their malignant transformation. The typical CAD system as shown in Figure 1, mainly includes: 1) colon segmentation, 2) 3D colon object reconstruction, 3) 3D centerline extraction as the optimal flying-path for navigation, 4) colonic polyp detection using either geometric features or statistical features, 5) color coding for polyp candidates visualization, 6) polyp features extraction, and 7) benign/malignant polyp classification. In this paper, we propose a general framework for colon segmentation and reconstruction. The proposed framework has been validated on 10 real CT colon datasets, and the segmentation accuracy has achieved 96.06%. The 3D colon objects are reconstructed from the segmented slices by computing their isosurfaces using MarchingCube algorithm.