First, the images (for example in DICOM, JPG or TIF format) are imported into the FXA software and a manual selection of the areas of interest is made.
The selection may be, for example, vertebrae, bony structures or implants.

The FXA software then processes the images with state of the art filters and procedures known from digital image processing. Hereby unwanted noise is reduced and the image clarity and definition is improved.

In the third step, each preselected object is overlaid between the various images and moved, scaled and rotated through a patent pending matching algorithm.
The main advantage of our proprietary algorithm is that the matching works independent of manual landmark placement, contour identification routines or a manual alignment between the images. For this reason, the analysis results of the FXA-method are operator independent and foster a reliable reproducibility. Furthermore, this algorithm is (to a certain level) robust against projection errors due to out-of-plane effects and can inherently compensate for these.

As a result of the matching process, the FXA software calculates translations, rotations and scaling factors and evaluates the differences amongst the objects of interest. Out of these differences, the relative motion can be calculated and quantified. The centers of rotation (CoR) can be determined utilizing the method described by Franz Reuleaux.

In the last step, the analysis results are processed in order to visualize the relative motion using animated images. An analysis report is prepared, which includes a statistical representation of the results, along with an assessment of the image quality and matching accuracy. The results are stored in a database and sent to the physician or CRO (online or by email) requesting the data.