Question

Attempting to set up a new image analysis tree using an image from a Fuji Oneshot phantom using IQWorks version 0.5

The fuji oneshot is a phantom containing a number of different objects including edges and distance rules.

Desired initial processes:

• MTF from the two edges included
• Scale measurement checks from the two measurement marks (the thin horizontal and vertical bars on the bottom and left of the image)
• Mean pixel values for the four regions of varying attenuation up the left hand of the image.

To begin with I selected: "Analysis -> Add Analysis -> MTF -> Edge/Line Analysis" from the menu bar. This placed a rectangular ROI on the image. On the pane to the left, under "process group 1" and linked in tree form appears "EdgeAnalysis2-Unnamed"

It is at this stage I remember that the image was obtained on a Fuji CR system and as such pixel value is not linearly linked to dose.

Could someone remind me of options to correct this then I will hopefully continue.

Environment

-- AlexSandison - 26 Oct 2009

Answer

If you answer a question - or have a question you asked answered by someone - please remember to edit the page and set the status to answered. The status is in a drop-down list below the edit box.

I have moved your question from the Documentation Web Alex to the support web.

-- EdMcDonagh - 26 Oct 2009

The first thing you need to do is find the STP (signal transfer properties) of the CR system, ie what pixel value you get at what dose. There are various sources for protocols for this, but a good start might be the KCARE ones.

You can then work out the relationship, and then change the signal calibrator (bottom left panel in IQWorks) to choose the correct relationship and enter the values you have calculated from the STP.

Don't choose the Fuji CR Antilog Calibrator as it does not function correctly.

I hope this helps.

-- EdMcDonagh - 26 Oct 2009

From KCare (MHRA report 04132) the equation is Pixel Value = -145.6 ln (incident air kerma (uGy)) + 256.2

-- AlexSandison - 28 Oct 2009

It has just been pointed out to me that you need version 0.5 or newer to use the signal calibrator function.

-- EdMcDonagh - 29 Oct 2009