Segmentation of Bone Tumor in MR Perfusion Images Using Neural Networks and Multiscale Pharmacokinetic Features

The decrease in the volume of viable tumor is an indicator for the effect preoperative chemotherapy has on bone tumors. We develop an approach for segmenting dynamic perfusionMR-images into viable tumor, nonviabletumor and healthy tissue. Two cascaded feed-forward neural networks are trained to perform the pixel-based segmentation. As features, we use parameters obtained from a pharmacokinetic model of the tissue perfusion (parametric images). Additional multiscale features that incorporate contextual informationare included. Experiments indicate that multiscale blurred versions of the parametric images together with a multiscale formulation of the local image entropy are the most discriminative features.

• Type: Conference
• Year: 2000
• Where: ICPR
• Authors: Michael Egmont-Petersen, Alejandro F. Frangi, Wiro J. Niessen, P. C. W. Hogendoorn, Johan L. Bloem, Max A. Viergever, Johan H. C. Reiber
• Keywords: Additional Multiscale Features, Computer Vision, ICPR 2000, Multiscale Blurred Versions, Multiscale Formulation