Artikel
Demonstration of an open source software module for two dimensional image registration and fusion
Vorstellung eines Open Source Software-Moduls zur zweidimensionalen Bildregistrierung und -fusion
Suche in Medline nach
Autoren
Veröffentlicht: | 30. Mai 2008 |
---|
Gliederung
Text
Objective: Aim of the project was the development of an open source based software module for intraoperative online image-fusion between preoperative MRI and intraoperative acquired ultrasound images. The fused images should facilitate the interpretation of the intraoperative ultrasound images. The plug-in is part of software able to run on a laptop and communicate with any DICOM ready ultrasound system via Ethernet port.
Methods: The developed plug-in is based on the open-source DICOM viewer OsiriX 2.7.5. The image registration is semi-automatic, anatomical landmark based. The neurosurgeon has to select two anatomical landmarks on each image and drag and drop the floating ultrasound image onto the reference MR image (a reconstructed slice from the preoperative MRI data). The floating image is translated, rotated and scaled, so that it can match the reference image, in a peer point matching fashion; each landmark of the first image is matched to the respective one of the second image. The registration and fusion of the two images is performed in less than three seconds.
Results: The output of the plug-in is a fused MR/ultrasound image. The underlying MR image is in grey scale, while a colour look-up table (CLUT) is applied to the overlaid ultrasound image. Image transparency can be adjusted by a slider, allowing the neurosurgeon to explore thoroughly the anatomy visualized by the two imaging modalities. The interpretation of the ultrasound image for tumour remands or artefacts, especially on the resection borders, is also facilitated. The registration quality depends solely on the accurate landmarks selection.
Conclusions: The developed open source software module facilitates the interpretation of the intraoperative ultrasound images by performing two dimensional image registration and fusion. In a next step the process will be augmented by an automatic landmark detection and image fusion scheme and the laptop-based system set-up will be integrated in the intraoperative neurosurgical workflow.