Topic 12: CT image formation + reconstruction Flashcards
(18 cards)
Define CT
- Computed tomography
- Volumetric imaging modality = based on X-ray absorption
- Allows reconstruction on 2D/3D
Describe the imaging of CT
- CT better than X-ray to show details/contrast of soft tissue
- CT lower spatial resolution than X-ray
- 1st modality = computer essential to image reconstruction
How long does it take a modern CT scanner to produce 2D image?
Less than a second
How much do CT scanners cost?
- Millions
- High cost per scan
Define data acquisition
- Collection of X-ray transmission measurement through patient
Describe data acquisition
- X-ray image projection = cannot determine exact location of area of interest
- Therefore radiologists take 2 perpendicular projections
- Requires X-ray source in shape of fan/cone
Describe the geometry for CT scanner + detectors
- Source + arc-shaped detector = rotate in tandem
- Projections = recorded through single plane within body for different angles
Factors reducing X-ray transmission
- Atomic composition of tissues
- Density of tissues
- Energy of photons
What happens to attenuated X-rays after passing through object?
- Collected by detector
- Converted from X-ray > electrical signals
- Signals converted > digital data = attenuation value calculated
What is the aim of CT?
- Get spatially resolved map of absorption coefficients = in 1 slice of body
- If map measured/sampled at small regular steps = finite resolution = can create digital image
Explain the mathematicas behind CT image reconstruction
- The object = 4 sectioned quadrant
- Initial beam = I₀ > gets weakened based on material it passes through
- Across = I₁ + I₂
- Down = I₄ + I₅
- Each block = absorption value = μ₁/μ₂/μ₄/μ₅ for each beam
- Once μ value is known = image can be reconstructed
What is the linear equation system used to solve for μ?
I = I₀ × exp(–μ × d)
Define Lambert-Beer’s law
- The output intensity is reduced depending on how much stuff (μ) the beam goes through over a certain distance (d)
Give the steps for image reconstruction
1) Measurements
2) Pre-processing
3) Raw data
4) Filtering
5) Filtered data
6) Interpolation
7) Back projection
8) Source images
Describe selected field of view
- Divided into small elements = pixels
- Pixels = make up cross-sectional image = represent small volume of tissue = voxel
- Typical tomographic axial image = 512x512/1024x1024 voxel = 260,000/1 mil greyscale each cross-section
Define density value of a pixel
- Depends on composition of tissue represented
- Units = Hounsfield = HU
How are HU calculated + range?
- From attenuation measurements relative to attenuation of H2O
- Range = -1024 > 3071
Give the HU values for different tissues
- Air = -1000
- Fat = -100
- Lung = -500
- Fat = -100 > -50
- Water = 0
- CFS = +15
- Blood = 30 > 45
- Muscle = 10 > 40
- Adrenal tumor = less than 10
- White matter = 20 > 30
- Grey matter = 37 > 45
- Liver = 40 > 60
- Bone = 700 > 3000
