Friedrich-Alexander-Universität
Friedrich-Alexander-Universität
Stanford Radiology

Minimal Scan

Minimal Scan Ranges for Non-Circular Objects

Application Example - Knees

The minimal scan range for circlular objects in fan-beam geometry is well known. Parker demonstrated a nice and efficient method to reconstruct such a short scan. Furthermore, a offset detector is able to double the FOV. However, this comes at the restriction that the scan range has to be increased to 360 degrees.

In this tutorial, we investigated the minimal scan range for non-circular objects. Here we only present a short example. The full research is found the Opens external link in new windowCT Meeting Paper by Herbst et. al.

The respective sources are found in package edu.stanford.rsl.tutorial.noncircularfov.

Complete Set Estimation

Knee Sinogram

In a first step, the object is projected into sinogram space and the image is thresholded to determine which rays have to be collected to create a complete reconstruction.

Scan Mask

scan configuration following the outline of the knees.

Then a mask is generated that represents the current scan configuration. Here, we chose to follow the outline of the knees from the left and chose a detector size of 250 pixels.

Sinogram Completion using Corresponding Rays

Completed mask

In a next step, the sinogram is completed using the information from redundant rays that are acquired in the fan-beam geometry. The image on the right shows the completed mask.

Missing Information

Missing rays

From the completed sinogram and the complete set, the missing rays can be determined as shown on the right.

Reconstruction

The present configuration is incomplete and resutls in artifacts in the image.

The missing information leads to artifacts in the reconstruction. The algorithm to find configurations that determine the minimal complete scan configuration is described in the above mentioned CT Meeting Paper.