Low complexity adaptive algorithms for Principal and Minor Component Analysis
Digital Signal Processing
| Hi-index | 0.00 |
Reliable segmentation of the colon is essential for automatic detection of polyps on virtual colonoscopy (VC). Large clinical trial suggests that providing patients with oral contrast agents may reduce the false positive rate of polyp detection even without bowel purgation, since residual fluid and remnants inside colon can be tagged out and become identifiable. Colon cleansing aims to remove the tagged materials from the computed tomography (CT) image, so that a completely air-filled virtual colon can be constructed. In this paper, we present an automatic colon cleansing method for virtual colonoscopy, which is composed of an automatic seed placement algorithm and a cell proliferation algorithm. Dataset containing 10 CT scans was used to test our method. The visual inspection demonstrated that the proposed method could fully automate colon segmentation for virtual colonoscopy without user intervention.