A flexible registration framework for multimodal image data

This work describes a registration framework based on the insight segmentation and registration toolkit (ITK) which can be used for matching multimodal image data. Different target groups with individual needs and precognition are addressed. The framework offers tools for supporting different matching tasks in a clinical environment. A setup editor defines specific rigid or non rigid matching approaches and the appropriate parameters. Different metrics including a correlation metric, a difference metric and mutual information based metrics are available. Furthermore, a test series editor can be used to evaluate the selected setup. The evaluation results, which are expressed in statistical figures, trends and performance measures, can be visualized and used for an optimal adapted setup configuration. Tests for matching precision, quality and parameter adjustments are offered. For export and import of image data, the most frequently used file formats of clinical environments like DICOM and ANALYZE are supported. We demonstrate some registration examples which frequently occur in the neurosurgical routine of a University Hospital.