Information Technology Reference
In-Depth Information
12.3
Fuzzy Images: Sources of Uncertainty
Despite of the variety of acquisition technologies, medical images suffer from low
quality in contrast with conventional photography, that produces relatively noise-
free images except where the grain of the film becomes visible. The main problems
of medical images are noise and partial volume effect. These phenomena suggest to
interpret them as fuzzy images.
12.3.1
Noise
All medical images contain some visual noise. The presence of noise gives an image
a mottled, grainy, textured, or snowy appearance. Image noise comes from a variety
of sources. No imaging method is free of noise, but noise is much more prevalent in
certain types of imaging procedures than in others.
Ultrasound imaging and nuclear images (PET, SPET) are generally the most
noisy. Noise is also significant in MRI, CT. In comparison to these, radiography
produces images with the least noise.
•
The noise in the PET image is mainly caused by attenuation of annihilation pho-
tons inside the body. Noise in photon counts are due to scattering and detection
of false positive coincidendes. Also the noise in PET images is highly dependent
on the reconstruction algorithm used.
•
In all imaging procedures using x-ray like CT, most of the image noise is pro-
duced by the random manner in which the photons are distributed within the
image. This is generally designated quantum noise. Recall that each individual
photon is a quantum (specific quantity) of energy. It is the quantum structure of
an x-ray beam that creates quantum noise.
•
In MRI the main source of noise in the image is the patient's body (radio fre-
cuency emission due to thermal motion). The whole measurement chain of the
MR scanner (coils, electronics...) also contributes to the noise. This noise cor-
rupts the signal coming from the transverse magnetization variations of the in-
tentionally excited spins.
•
The noise presented in ultrasound images is due to the variation in the detector
sensitivity, environmental variations, the discrete nature of radiation, transmis-
sion or quantization errors, etc. This produces speckle noise and weak edges
which makes difficult to identify the regions of interest.
12.3.2
Partial Volume Effect
The term partial volume effect (PVE) actually refers to 2 distinct phenomena that
make intensity values in medical images differ from what they ideally should
be [32].
The first effect is the 3-dimensional image blurring introduced by the finite spa-
tial resolution of the imaging system. For example, the spatial resolution in PET
images is limited by the detector design and by the reconstruction process.
The
