| 2. |
- Sandborg, Michael, et al.
(author)
-
Implementation of unsharpness and noise into the model of the imaging system : Applications to chest and lumbar spine screen-film radiography
- 1999
-
Reports (other academic/artistic)abstract
- A model of the complete x-ray imaging system including the patient is a powerful tool for imaging system analysis and the optimisation of image quality and patient dose. It allows flexible variation of the system components (i.e. x-ray source, antiscatter device and image detector) and study of their effect on image quality and patient risk. Our group has developed, validated and calibrated a Monte Carlo model of the complete imaging system for chest and lumbar spine examination including voxalised human male anatomy. The Monte Carlo program calculates the contrast and signal-to-noise ratio (SNR) of various contrasting details within the voxel phantom. Important details in the images have been selected by consulting radiologist and the EU document of image quality criteria. The entrance surface dose without back-scatter and the effective dose are used as measures of patient radiation risk. The contrasts of the details are derived initially from Monte Carlo estimates of the energy imparted per unit area to the image detector beside and behind the detail. However, this ignores the effects of unsharpness in the imaging chain (such as screen-film, geometric and motion unsharpness) and the influence on contrast of the film characteristic curve. In the Monte Carlo program, SNR is calculated assuming that the noise arises from the random fluctuations in the energy imparted per unit area to the image detector only. However, other noise sources also contribute to the total noise, such as screen and film noise. Hence the model of the imaging system needs to be further developed to take these effects into account. The methods used to extend the model are described below together with illustrations of their effect on the difference in optical density, DOD, and SNR in chest and lumbar spine imaging.
|
|
