Registration Server

OEM customers and research institutions are now taking advantage of Mirada’s registration technologies through a “black box” server side solution. Mirada silently provides clinical quality registration results to support numerous clinical applications. Some examples of these are:

Anatomical matching: this basic use of registration information simply permits the user to co-locate similar anatomical points in different images of the same patient and body part. The typical implementation uses the registration server to generate a rigid transformation matrix or non-rigid deformation field, and then uses this implementation to co-locate a pointer or crosshair which has been placed by the user in one image with the same anatomical location in the second image.

Region tracking: this is an extension of anatomical matching. A region is defined in one image, and then the registration server is used to locate the same region in subsequent images. It is possible to track from the first image through multiple subsequent images.

Subtraction: sometimes called difference imaging, this application first registers two examinations of the same modality and then quite simply subtracts the intensities from one image from the other to show the difference. A typical example of this would be pre and post contrast CT imaging, where subtraction may be used to leave only the contrast enhanced regions in the final picture

Atlas based segmentation: an anatomical atlas may be built, usually by averaging multiple previous similar images (for example, whole body CTs, or PET brain scans) and then using an expert clinician to label anatomical regions on the averaged dataset. Successive future patient scans may then be registered to the average image and the anatomical labels transferred to the current scan of interest.

Comparison: this is generally achieved using a multiple volume renderer (such as those found in Mirada's XD3 product) to allow two fused images to be viewed in the same space. Various techniques may be used to support the comparison such as allowing the user to blend one image into the other, allowing the user to switch quickly from one image to the other (which has the affect of highlighting differences) or other rendering techniques such as checkerboard overlays and spy-glass windows (which allow the user to look "through"one image into the other)