FRCR physics notes contents

Free mock FRCR anatomy exams

18 mock anatomy exams for the first FRCR examination

Join 10k+ newsletter subscribers

Please note: Your email address will never be shared with any 3rd parties. It will only be used for Radiology Cafe communications. Emails are sent less than once a month on average. Read our Privacy policy for more details.

Image quality

There are certain qualities of an image that affect each other and determine the quality of the displayed image:

  1. Contrast
  2. Resolution
  3. Noise

As well as:

  1. Unsharpness
  2. Magnification
  3. Distortion
  4. Artefacts

1. Contrast

Contrast is the difference in the displayed or image signal intensity between two areas of interest e.g. a lesion and background tissue. A high contrast image has a greater difference between the grey shades displayed but a smaller range of greys. A low contrast image has a smaller difference (i.e. it’s more difficult to make out different areas) but a larger range of greys.

High and low contrast
High and low contrast

Subject contrast

Subject contrast is the ratio of the radiation intensities in different parts of an image due to the quality of the subject being imaged. The contrast is due to the differential attenuation by the tissues.

c ∝ (μ1 – μ2) x t


c = contrast
μ = attenuation coefficient of object 1 and 2 in the material being imaged
t = thickness of the structure

From the above equation you can see that a higher contrast is achieved with:

  • Thicker structure being imaged
  • Greater difference between the attenuation of the two objects
Subject contrast
Subject contrast

In the diagram tissue A absorbs 50% of the radiation incident upon it, B absorbs 90%. If there are 1000 photons for every element of the image then 500 photons will emerge from A and 100 from B (a ratio of 5:1).

As optical densities (the displayed shade in the image) vary with the log of the exposure log500 = 2.7 and log100 = 2.0 so the subject contrast has a difference in the logs of 0.7.

Factors affecting contrast

Linear attenuation coefficient of subject

The linear attenuation coefficient depends on the Compton and the photoelectric linear attenuation coefficient (LAC).

Compton LAC = ρ / E

Photoelectric LAC = ρZ3 / E3


ρ = density
E = energy (kV)
Z = atomic number of material

From the equations above we can see contrast can be improved by:

  • Decreasing the energy (tube potential kV)
  • Increasing the difference in Z (atomic number) (e.g. use of iodine or barium as a contrast medium against soft tissue)
  • Increasing the difference in ρ (density) (e.g. use of barium or gas as a contrast medium)
Overlying tissue

If there is overlying tissue over both A and B, subject contrast is not changed as the same ratio of photons is still absorbed in tissues A and B.


Suppose scatter contributes an additional 50 photons to each element in the image. There will now be 550 photons in the film under tissue A and 150 under tissue B. The ratio of signals is now 3.6 (550/150) and the difference in logs is 0.6 (was 0.7) i.e. a reduced contrast.

Scatter is reduced by:

  • Using an anti-scatter grid
  • Using a larger air gap


Improved contrast
  • Thicker structure
  • Greater attenuation between objects
  • Decreasing kV
  • Increasing difference in Z of objects
  • Increasing difference in density of objects
Reduced contrast
  • Increased scatter
No effect
  • Overlying tissue

Image contrast

Image contrast, or radiographic contrast, is the difference in density between neighbouring regions on the image.

Digital windowing
Digital windowing

Image contrast is altered by windowing on the viewing monitor. Images are presented at a certain width and centre of Hounsfield units displayed. The larger the width, the larger the range of shades displayed and, therefore, the smaller the difference in contrast between each shade. The window is adjusted for the Hounsfield unit of the tissues that need to be assessed.

Narrow and wide windows
Narrow and wide windows

Written by radiologists, for radiologists with plenty of easy-to-follow diagrams to explain complicated concepts. An excellent resource for radiology physics revision.

2. Spatial resolution

Resolution is the measure of how far apart two objects must be before they can be seen as separate details in the image. There are several ways to measure spatial resolution.

Measuring spatial resolution

Line spread function

This is a measure of how spread out the image of a sharp object becomes. However, this is difficult to calculate and it is easier to look at the image in terms of spatial frequency content.

Spatial frequency

Spatial frequency in line pairs per mm
Spatial frequency in line pairs per mm

This is measured in line pairs per mm (lp/mm). An image with a high lp/mm is a high spatial frequency image as there are many alternating light and dark regions in a single millimetre. We, therefore, need a system that can reproduce the image with the appropriate frequency. The lp/mm of different radiographic techniques can be found in the appendix.

How well a system is able to represent the object spatial frequency is expressed as the modulation transfer function (MTF).

Modulation transfer function
MTF = 1Same range is obtained in the image
MTF < 1Lower range in the image
MTF = 0No information in the image

E.g for an imaging system that can fully change from black to white over 1 mm:
For images with 0.5 lp/mm, it gives an MTF of 1
For images with > 0.5 lp/mm, it gives an MTF of <1

MTF is calculated from the line spread function using Fourier transform analysis. The total MTF is the product of the MTF of all constituent parts of the imaging system.

Factors affecting spatial resolution

System spatial resolution
System spatial resolution
  • If the object spatial frequency is too high for the system, the system will be unable to display the image adequately. The higher the object spatial frequency, the lower the MTF until the system cannot distinguish the line pairs at all resulting in a homogeneous grey i.e. MTF = 0.
  • If the object has low contrast the system will reach an MTF of 0 earlier as the smaller difference in the range of shades means that the image will reach a homogeneous grey much sooner than if it was a high contrast image (e.g. alternating bands of black and white).
  • Anything that increases the unsharpness will blur the edges and further reduce the spatial frequency.

Digital detectors

There are several things that affect the resolution of digital detectors.

Detector properties
Detector properties

Detector aperture

This signal is averaged over the area of the detector element. If object details are much smaller than the size of the element they are not visible unless they have enough contrast to have a significant effect on the average signal.

Sampling pitch

This is the centre-to-centre distance between individual detector elements. It determines the highest spatial frequency that can be imaged: the Nyquist frequency.

Nyquist criterion states that the sampling frequency must be at least twice the highest signal frequency. The highest signal frequency is also called the “Nyquist frequency” i.e. for a system to be able to accurately represent the spatial resolution of the object it must have the appropriate sampling pitch which is no less than double the object spatial frequency.

Sampling frequency = 2 x Nyquist frequency

Sampling frequency limit
Sampling frequency limit

3. Noise

There is random variation in the number of photons forming each part of the image, called noise, that can obscure the signal received from the subject. The amount of quantum noise produced increases with an increasing total number of photons. We usually express this random variation as the standard deviation which is best estimated by the square root of the average number of photons per area.

Quantum noise ∝ √photons

However, when we calculate the quantum noise as a proportion of the total signal we can see that the proportion of noise in the signal actually decreases with an increasing photon concentration and:

Noise ∝ 1/√photons

Average number of photons absorbed by each detector (N)1000100
Noise√1000 =
√100 =
Proportion of signal which is noise31/1000 x 100 =
10/100 x 100 =
Signal to noise ratio (SNR) = N / √N3110

Reducing the proportion of noise in an image will improve the quality. The main way to achieve this is to increase the number of photons detected and used to form each image pixel / element. This can be done in several ways.

  • Increasing the dose (mA): higher number of photons and smaller proportion of noise
  • Using an image receptor with a greater attenuation coefficient: more photons are absorbed and converted into a signal
  • Make the image receptor thicker: again, more photons will be absorbed and converted into a signal
  • Using larger detector elements: more area to absorb photons per pixel. However, the spatial resolution will decrease

Factors that don’t reduce noise

  • Amplification: attaining a higher signal from each absorbed photon, either by using a faster film-screen combination or gain of an image intensifier would just amplify the signal from noise as well
  • Using a narrower window to produce a high contrast image

4. Unsharpness

There are four causes of unsharpness:

  1. Geometric unsharpness
  2. Image receptor unsharpness
  3. Movement unsharpness
  4. Edge unsharpness

a. Geometric unsharpness

Geometric unsharpness
Geometric unsharpness

The boundaries between a dark and a light area may be ill-defined, resulting in a blurred edge. This is called “unsharpness”. There are several causes and types of unsharpness as outlined below.

The focal spot is not infinitely small. There will be areas of the image that are:

  • High signal: all x-ray photons reach detector
  • Low signal: no x-ray photons have passed through the object to reach the detector
  • Intermediate: not all photons have passed through the object. The size of this area determines the unsharpness and is called the penumbra.

Moving an object closer to the focal spot will increase the penumbra and, therefore, the unsharpness.

The geometric unsharpness (Ug) is determined as follows:

Ug = f x b / a
f = x-ray focal spot size
a = distance from x-ray source to front surface of object
b = distance from object to detector

b. Image receptor unsharpness

  • Digital images: if a detector element lies across the border between a light and a dark area the pixel displayed will be an average of these two values creating a blurred border.

c. Movement unsharpness

If an object moves during the acquisition the edge will be blurred resulting in unsharpness.

d. Edge unsharpness

Edge unsharpness
Edge unsharpness

If an object has a tapering edge the attenuation will gradually decrease along the object.

5. Magnification

Magnification artefact
Magnification artefact

Magnification (M) depends on the relative distance of the object between the x-ray source (focal spot) and the image receptor. The further from the detector the object is the more the image is magnified.

M = image size / object size

= d2 / d1

6. Distortion

Depending on the angle at which the x-ray beam passes through an object it can distort the shape and create a distortion artefact.

7. Artefacts

There are a variety of patient and system factors that can create artefacts:

  • Motion artefact
  • Double exposure
  • Grid cut off
  • Radio-opaque objects on or external to the patient

Σ  Summary


  • Difference in attenuation (subject contrast) or displayed shade (image contrast) of an image

Subject contrast

  • contrast proportional to (μ1 – μ2) x t – where: μ = attenuation coefficient of object 1 and 2, t = object thickness
  • Improve contrast by:
    • Thicker object
    • Greater attenuation between objects
    • Decreasing kV
    • Increasing Z (atomic number) difference in objects
    • Increasing difference in density of objects
  • Contrast reduced by:
    • Increased scatter (no anti-scatter grid, smaller air-gap used)
  • No effect:
    • Overlying tissue

Image contrast

  • Digital imaging system: achieved by windowing at the imaging monitor

Spatial resolution

  • Measure of how far apart two objects must be before they can be seen as separate details in an image

Measured as:

  • Line spread function: how spread an image of a sharp object becomes. Difficult to measure
  • Line pairs per mm (lp/mm)

Accuracy of system display of object spatial frequency = modulation transfer function (MTF)

  • MTF = 1 same range obtained
  • MTF < 1 smaller ranger obtained
  • MTF = 0 no information in image

Factors affecting spatial resolution:

  • Object properties:
    • Object spatial frequency: if it is too high for the system it will not be displayed accurately 
    • Object contrast: lower contrast objects reach an MTF of 0 at lower spatial frequencies
  • Computed / digital radiography:
    • Detector element size: smaller element = higher spatial resolution
    • Distance between detector elements: smaller distance = higher spatial resolution
  • Others:
    • Anything that increases unsharpness


  • Noise inversely proportional to √photons
  • Reducing noise: anything that increases number of x-ray photons (x-ray beam) produced and absorbed and the number of light photons (at the image receptor) produced:
    • Increasing dose (mA)
    • Using an image receptor with a greater attenuation coefficient
    • Making the image receptor thicker
    • Using larger detector elements

Factors that don’t reduce noise:

  • Amplification
  • Using narrower window to produce a high contrast image



  • Geometric unsharpness
    • Focal spot not infinitely small, therefore, blurred penumbra produced at object edge
    • Ug = f x b / a – where f = x-ray focal spot size, a = distance from focal spot to object, b = distance from object to detector
  • Image receptor unsharpness
    • Digital images: if detector element straddles light and dark area pixel displayed will be an average of these two values
  • Movement unsharpness
  • Edge unsharpness


  • Greater magnification by moving object further from detector and closer to x-ray tube


  • Due to the finite size of the focal spot an image may be distorted depending on the angle at which it is imaged


  • These may be due to patient or system factors

Digital radiography – special notes


  • Speed = 2000 / x where x = dose incident on IP
  • S < 200 = improved SNR but increased patient doses

Spatial resolution

  • Measured by MTF
  • Detective quantum efficiency (DQE) = SNR2out/SNR2in
    • Greater DQE = more efficient detection of incident x-ray photons
  • Improved by:
    • Smaller diameter of readout laser beam
    • Smaller pixels
    • Smaller size of phosphor crystals
    • Thinner phosphor layer
    • No light reflection / absorption backing layer (produces scatter)