Measurement of Contrast Resolution
24.11.2019, 18:20
While unsharpness shows the resolution perpendicular to the x-ray beam, contrast defines the resolution in direction of the beam.
ASTM E1316 (Definitions) defines contrast sensitivity as "a measure of the minimum percentage change in an object which produces a perceptible density/brightness change in the radiological image". As far as I know there is only one standardized IQI for measurement of the contrast independent of the imaging system spatial resolution limitations: ASTM E1647 with recesses large enough to exclude unsharpness issues:
The blocks with the recesses are available in eight different material types and four different sizes. Here is a picture of four of them:
All other IQIs measure contrast and spatial resolution at the same time with the IQI, e.g. hole penetrameter, step-hole penetrameter or wires.
ASTM E1647 (-94: Standard Practice for Determining Contrast Sensitivity in Radioscopy) is one of the radiocopy standards but since 2003 renamed to Standard Practice forDetermining Contrast Sensitivity in Radiology and also included in the CR Phantom of ASTM E2445 in the versions of AL, CU and SS.
The evaluation is quite easy: You look at the image and the shallowest recess that can be repeatably and reliably imaged shall determine the limiting contrast sensitivity in [%].
If you need repeatable results you can measure in ROIs with about half of the size of the recess inside the recesses the mean pixel value and substract the mean pixel value of an similar size ROI beside the recess. The result is the absolut contrast of the percentage of the recess. E.g. if you measure 200 PV difference of the 2% recess you have 100 PV for 1% material difference. Alternativly you also could use a line profile function across the four recesses and read the contrast at the scale on the left in absolute PVs.
Normally there is always a 1% contrast in the image measurable but sometimes you can not see the 1% recess. This may happend due to noise in the image. Therefore normally all requiremends are defined in Contrast to Noise Ratio (CNR).
Large object (like the recesses) are visible in the image if their signal is larger than the noise level.
Hint: For small objects - like the holes of the E1025 penetrameters - a CNR of 2.5 is required; this value was valid in the aerospace intustry for years because of a study which was done with a 127µm detector. Different SRBs would require different factors - maily smaller for smaller SRBs.
ASTM E1316 (Definitions) defines contrast sensitivity as "a measure of the minimum percentage change in an object which produces a perceptible density/brightness change in the radiological image". As far as I know there is only one standardized IQI for measurement of the contrast independent of the imaging system spatial resolution limitations: ASTM E1647 with recesses large enough to exclude unsharpness issues:
The blocks with the recesses are available in eight different material types and four different sizes. Here is a picture of four of them:
All other IQIs measure contrast and spatial resolution at the same time with the IQI, e.g. hole penetrameter, step-hole penetrameter or wires.
ASTM E1647 (-94: Standard Practice for Determining Contrast Sensitivity in Radioscopy) is one of the radiocopy standards but since 2003 renamed to Standard Practice forDetermining Contrast Sensitivity in Radiology and also included in the CR Phantom of ASTM E2445 in the versions of AL, CU and SS.
The evaluation is quite easy: You look at the image and the shallowest recess that can be repeatably and reliably imaged shall determine the limiting contrast sensitivity in [%].
If you need repeatable results you can measure in ROIs with about half of the size of the recess inside the recesses the mean pixel value and substract the mean pixel value of an similar size ROI beside the recess. The result is the absolut contrast of the percentage of the recess. E.g. if you measure 200 PV difference of the 2% recess you have 100 PV for 1% material difference. Alternativly you also could use a line profile function across the four recesses and read the contrast at the scale on the left in absolute PVs.
Normally there is always a 1% contrast in the image measurable but sometimes you can not see the 1% recess. This may happend due to noise in the image. Therefore normally all requiremends are defined in Contrast to Noise Ratio (CNR).
Large object (like the recesses) are visible in the image if their signal is larger than the noise level.
Hint: For small objects - like the holes of the E1025 penetrameters - a CNR of 2.5 is required; this value was valid in the aerospace intustry for years because of a study which was done with a 127µm detector. Different SRBs would require different factors - maily smaller for smaller SRBs.
Re: Measurement of Contrast Resolution
24.11.2019, 19:08
There is one other IQI which can be used to measure the contrast. Due to the high required dynamic range of the detector this IQI is used with DDAs only.
The grooved stepwedge from ASTM E2597:
The intention of this IQI is to measure the Contrast Resolution of DDAs within a larger material thickness range.
The benefit of the different steps are the visible dynamic range.
If the applications requires a SNR of >250 this detector could be used with an accumulation time of 16s up to 78mm of Aluminum.
Discussion: The material AL 6061 is not very smooth; there are a lot on internal structures which limits the archivable CS due to the noise of the internal structures. I would prefer a more smooth material like AL7022 - where the SNR could be >1500. Please post your opinions.
The grooved stepwedge from ASTM E2597:
The intention of this IQI is to measure the Contrast Resolution of DDAs within a larger material thickness range.
The benefit of the different steps are the visible dynamic range.
If the applications requires a SNR of >250 this detector could be used with an accumulation time of 16s up to 78mm of Aluminum.
Discussion: The material AL 6061 is not very smooth; there are a lot on internal structures which limits the archivable CS due to the noise of the internal structures. I would prefer a more smooth material like AL7022 - where the SNR could be >1500. Please post your opinions.
Hosted by iphpbb3.com
Impressum | Datenschutz