Thin layer tissue classification for electronic cleansing of CT colonography data

CT colonography (CTC) is a rapidly evolving technique to screen for colorectal polyps. Fecal residue may occlude or, reversely, mimic polyps. Electronic cleansing aims at removing contrast-enhanced fecal residue from the image. However, thin layers of soft tissue (the colon wall or a fold) or residue are easily misclassified by current electronic cleansing methods, thereby causing holes in the colon wall or other artefacts that hamper visualization and automated detection. We present a thin layer model to detect and characterize such layers to support electronic cleansing. It is demonstrated that the model sustains robust estimation of the location and thickness of such a layer. Such thicknesses of thin layers were measured in real data sets. A lower bound on the thickness of such layers exists and was found to

• Type: Conference
• Year: 2008
• Where: ICPR
• Authors: Vincent Frans van Ravesteijn, Frans M. Vos, Iwo Serlie, Roel Truyen, Lucas J. van Vliet
• Keywords: Computer Vision, Contrast-enhanced Fecal Residue, Electronic Cleansing, Fecal Residue, ICPR 2008