Differences in the pixel values of a homogeneous object. The result of these fluctuations is an image that is grainy.
Noise degrades image quality and affects the perceptibility of detail.
Noise can be measured by observing the standard deviation of a given set of pixels within an ROI.
There tends to be three major sources of noise within a CT image. These include:
Quantum mottle- Quantum mottle is the result of an x-ray photon flux. When there are not enough photons that strike the detector, it results in an image that appears noisy when reconstructed.
Component limitations– Limitations within the physical components of the CT scanner can also attribute to the noise of an image. This may also be called electronic noise.
Reconstruction algorithms– The reconstruction algorithm that is used for a particular set of images can also impact noise. An algorithm that reconstructs an image with high resolution also tends to be an image with increased noise.
The signal-to-noise ratio is what is used to quantify the amount of noise in a CT image. It is important to note that an improvement in the signal-to-noise ratio will also improve contrast resolution, but it often results in decreasing the spatial resolution.
The same is true for increasing spatial resolution; it often results in a decrease in signal-to-noise ratio and contrast resolution.
Factors that impact noise are:
Photon Flux– Photon flux is the rate that photons pass through an area in a set of time. Increasing the rate of photons results in a decrease in noise. Increasing mA, kVp, or the scan time can also increase the photon flux within an image, but this comes at the cost of increased patient dose and decreased contrast resolution.
Pitch– A high pitch results in a noisy image because less photons are interacting with the detectors. In order to utilize a higher pitch without sacrificing image quality, the dose must be increased.
Patient– The larger the patient, the noisier the image will be unless the radiation dose is adjusted to compensate. Dose modulation can be used to adjust the patient dose to produce an image with an acceptable amount of noise, while keeping the increased patient dose to a minimum.
Algorithm– As mentioned previously, the use of reconstruction algorithms that require high resolution result in an image with increased noise. An example of a high resolution algorithm would be the bone window.
Voxel- The size of the voxel has a direct impact on the level of noise within an image. As the size of the voxel increases, the noise decreases. However, as voxel and pixel size increases it results in a loss of detail and lower spatial resolution.