To be efficient, inspired, and informed.
When choosing the right CT scanner for your sample or research, the resolution of the instrument is one of the most important specifications to check. However, this is one of the most elusive things to figure out. Why is that?
Voxel size and resolution are terms often used interchangeably in X-ray CT, but they’re not the same thing. This is a big part of the reason why the resolution of a given CT scanner is hard to figure out.
Let’s quickly review what each term means and why the distinction matters.
What is voxel size?
Voxel size is the 3D equivalent of pixel size. It’s determined by the detector pixel size, the scan geometry (source-to-detector and source-to-object distances and lens magnification when a lens is used), and reconstruction settings. A smaller voxel size means the object is sampled more finely, but that alone doesn’t guarantee higher resolution.
Additionally, when selecting a CT scanner, note that the voxel size depends on the FOV (field of view). Large FOV limits the achievable voxel size.
What is resolution?
Resolution is the ability of the system to distinguish two separate features. They could be two objects close to each other or one small thing and the background. The voxel size needs to be half or smaller than the feature size in a cross-section to see that it is there. The smallest thing you can see practically defines the spatial resolution, and it is always more than double the voxel size.
Furthermore, while voxel size is rather simply defined by the magnification factor, detector pixel size, and reconstruction settings, resolution is affected by many factors, such as X-ray source spot size, point spread function of the detector, mechanical stability of the scanner and the sample, contrast, signal-to-noise ratio, and reconstruction algorithms. This is why reducing voxel size alone doesn’t always improve effective resolution.
A practical rule of thumb:
For general imaging, the achievable resolution (spatial resolution) is typically about two times larger than the voxel size. Using a voxel size of 1/5 to 1/10 of the size of the feature you want to resolve is recommended.
In metrology applications, specialized surface determination methods are often used to achieve higher resolution from a limited voxel size. For example, using locally adjusted ISO-50 surface determination, dimensional accuracy can reach about 1/10 of the voxel size.
How to think about it:
To evaluate a CT scanner's resolution, pick a representative sample and define the FOV. Check the smallest voxel size achievable on the machine by magnification (not by reconstruction) for that FOV. Multiply that number by 2. That is the spatial resolution. Multiply that voxel size by 5 to 10. That is the smallest feature or "thing" you can see.
To learn more, you can check out the blog articles and on-demand webinars below:
