PMOD Image Matching and Fusion Tool Introduction

The PMOD image registration and fusion tool (PFUS) supports a large range of tasks related to image matching and fusion such as

• the manual realignment of volumetric image sets,
• automatic image registration using different methodologies,
• SPM-type spatial normalization (e.g. to stereotactic template images),
• motion correction of dynamic series,
• a multitude of fused renderings of registered images,
• saving of fused representations in JPEG/TIFF or as DICOM secondary capture images,
• pixel-wise image algebra on registered images,
• scatter plots of VOI pixel values in 2 (2D plot) or 3 (3D plot, optional) matched studies,
• volume-of-interest definition directly in the fused image display.

Overview of the PFUS Operation Principle

The PFUS tool regards one study as the Reference. All other studies are adjusted in pixel size and slice thickness to the reference study. This is done by interpolating their image information within oblique planes across the image volume, a process called reslicing. Hence the images to be adjusted are called the Reslice studies in this text. As a result of reslicing, the Reference and the Reslice studies have identical resolution, and fusion images can be easily generated by combining the pixel values in the different series. For instance, the color values can be added (blending mode), or only one of the image values can be shown depending on a threshold value (overlay mode).

If the anatomic structures in the Reference and the Reslice studies are already in spatial agreement, only a mere resolution adjustment of the Reslice study is required, for example to interpolate a 128x128 PET to the 512x512 matrix of the CT (Reference) in a combined PET/CT study. In other situations, however, an adjustment of the geometry is also required to bring the anatomy in both studies into agreement. For example, in most cases it is required to rotate and shift the images of a brain PET study to match them with the images of a MRI study, and interpolate them to the 256 x256 MRI matrix.

After the reslicing/matching step, all image sets have the same resolution. As a consequence, all of them can be used for fusion purposes, pixel-wise arithmetics as with a "hand-held calculator", and also for the evaluation of VOI pixel values with scatter plots.

Data Size Considerations

As the Reference determines the final image resolution, it is recommended to use the higher resolved image as the Reference for avoiding losses in image quality. The user, however, should be aware that the size of the Reslice study may increase dramatically. For instance, if a dynamic PET study is matched to a 256x256 MRI with thin slices, the size of the resliced PET data can easily grow by a factor of 10. Such big data sets can become a problem for the available RAM, and for subsequent processing steps.

Starting the Fusion Tool

The image fusion tool PFUS is started with the Fusion button from the PMOD ToolBox

or by directly dragging image files from the desktop onto the above button. PFUS organizes the available tasks on different pages which are explained below.

Basically, one page is for database loading only, two pages are for matching purposes, while the last two provide advanced fusion displays.

Taskbar

The taskbar on the right side of the application window provides a quick shortcut for several frequent tasks. Please note the tooltips which provide short explanation of the button functionality.

Further Information

Please note that the following description is intended as a reference and not as a tutorial. For practical examples how to work with the PFUS software please refer to the PMOD Workbook, which is also distributed as part of the PMOD documentation.