Digital Radiography
Detector Correction in Digital Radiography
There are 3 detector corrections used in digital radiography; dead pixel correction, flat field correction and dark noise correction.
Dead Pixel Correction
Sometimes detector elements malfunction due to radiation damage or radiation exposure, making the pixel appearing as either white or black in the image. This is corrected by averaging the surrounding pixel values and giving this resultant value to to the 'dead pixel'.
Flat Field Correction
Sometimes there are subtle differences in the sensitivity of each detector element, and the gain associated with amplifiers connected to individual elements varies. This means there can be a varying output signal for the same exposure. This is corrected by taking a 'flat field' or uniform image, which is then used to compensate for differences in sensitivities and gain for detectors. Following flat field correction, a uniform signal will be achieved.
Dark Noise Correction
Dark noise correction must be done when electronic noise causes detector elements to produce a signal in the absence of x-rays. This results in a finite grey scale without any exposure. To correct this, a series of dark images are taken and the average is then subtracted from the raw image of the patient.
Windowing and Levelling
When viewing an image, we differentiate between tissues using a grey scale value. This grey scale can be manipulated to encompass all pixel values or a portion of pixel values, using the process of windowing. Through levelling, we select a pixel values around which the window, or range, will be surrounded. Selecting the level determines brightness. To enhance the dark portion of the image, a low window is selected, with any structure above appearing as bright white. To enhance the contrast in the bright portion of the image, a higher window and level are selected, with any structure below the window appearing as black.
Look Up Tables or LUTs are sets of data installed in the computer to automatically adjust windows and levels for a certain image by substituting in-built pixel values for those in the image.
As a general rule, windowing enhances contrast. The smaller the window, the greater the contrast.
Convolution
Image post-processing often involves convolving the pixel matrix with a kernel. Convolution is the combining of the image matrix with another matrix, known as a kernel. Different kernels can be applied to cause specific image effects, e.g. sharpening, blurring and edge detection. This can be expressed using the following equation:
where G and H are functions of x and k is the kernel. By applying a convolution to a digital image, grey scale values are shifted and added.
DICOM
DICOM stands for Digital Imaging and Communication in Medicine and is a standard image format used for handling, storing and transporting information in medical imaging. Special software is required to view the images in DICOM format.
Contrast Agents
Many anatomical structures have similar anatomical numbers and densities, making it difficult to distinguish on plain x-ray. When wanting to view a specific structure, we may introduce radiographic contrast agents. Contrast agents have high atomic numbers and highly attenuate x-rays, momentarily changing the atomic number of the structure in which it accumulates. The purpose of contrast is to enhance the appearance of the anatomy of interest.
Iodine and barium are two commonly used contrast agents. Iodine has an atomic number of 53 and barium 56. Iodine is commonly used in angiography to improve vessel appearance, as well as renal, gallbladder, urinary tract and bile duct imaging. Barium is mainly used for enhancing gastrointestinal structures. Barium is introduced by enema or meal.
Digital Subtraction Angiography
Digital subtraction angiography is an application of digital fluoroscopy which employs a rotational, biplane C-arm configuration with a flat panel detector. DSA is used mainly to visual blood vessels, for example, aneurysms. In DSA, a mask scan is undergone, which is a normal fluoroscopic scan of the area of interest. Iodine contrast is then injected via a catheter and a second 'fill' scan is undergone. The two scans are identical, except the mask does not contain any contrast. As they are otherwise identical, the mask scan can be logarithmically subtracted from the fill scan to remove unnecessary background structures:
Spatial Resolution
Spatial resolution is the size of the smallest object than can be resolved by the imaging system. High spatial resolution improves image quality. Spatial resolution can be tested using a phantom which a series of lines which indicate from low to high resolution. Spatial resolution is affected by:
There are 3 detector corrections used in digital radiography; dead pixel correction, flat field correction and dark noise correction.
Dead Pixel Correction
Sometimes detector elements malfunction due to radiation damage or radiation exposure, making the pixel appearing as either white or black in the image. This is corrected by averaging the surrounding pixel values and giving this resultant value to to the 'dead pixel'.
Flat Field Correction
Sometimes there are subtle differences in the sensitivity of each detector element, and the gain associated with amplifiers connected to individual elements varies. This means there can be a varying output signal for the same exposure. This is corrected by taking a 'flat field' or uniform image, which is then used to compensate for differences in sensitivities and gain for detectors. Following flat field correction, a uniform signal will be achieved.
Dark Noise Correction
Dark noise correction must be done when electronic noise causes detector elements to produce a signal in the absence of x-rays. This results in a finite grey scale without any exposure. To correct this, a series of dark images are taken and the average is then subtracted from the raw image of the patient.
Windowing and Levelling
When viewing an image, we differentiate between tissues using a grey scale value. This grey scale can be manipulated to encompass all pixel values or a portion of pixel values, using the process of windowing. Through levelling, we select a pixel values around which the window, or range, will be surrounded. Selecting the level determines brightness. To enhance the dark portion of the image, a low window is selected, with any structure above appearing as bright white. To enhance the contrast in the bright portion of the image, a higher window and level are selected, with any structure below the window appearing as black.
Look Up Tables or LUTs are sets of data installed in the computer to automatically adjust windows and levels for a certain image by substituting in-built pixel values for those in the image.
As a general rule, windowing enhances contrast. The smaller the window, the greater the contrast.
Convolution
Image post-processing often involves convolving the pixel matrix with a kernel. Convolution is the combining of the image matrix with another matrix, known as a kernel. Different kernels can be applied to cause specific image effects, e.g. sharpening, blurring and edge detection. This can be expressed using the following equation:
G(x) = H(x) x k(x)
where G and H are functions of x and k is the kernel. By applying a convolution to a digital image, grey scale values are shifted and added.
DICOM
DICOM stands for Digital Imaging and Communication in Medicine and is a standard image format used for handling, storing and transporting information in medical imaging. Special software is required to view the images in DICOM format.
Contrast Agents
Many anatomical structures have similar anatomical numbers and densities, making it difficult to distinguish on plain x-ray. When wanting to view a specific structure, we may introduce radiographic contrast agents. Contrast agents have high atomic numbers and highly attenuate x-rays, momentarily changing the atomic number of the structure in which it accumulates. The purpose of contrast is to enhance the appearance of the anatomy of interest.
Iodine and barium are two commonly used contrast agents. Iodine has an atomic number of 53 and barium 56. Iodine is commonly used in angiography to improve vessel appearance, as well as renal, gallbladder, urinary tract and bile duct imaging. Barium is mainly used for enhancing gastrointestinal structures. Barium is introduced by enema or meal.
Digital Subtraction Angiography
Digital subtraction angiography is an application of digital fluoroscopy which employs a rotational, biplane C-arm configuration with a flat panel detector. DSA is used mainly to visual blood vessels, for example, aneurysms. In DSA, a mask scan is undergone, which is a normal fluoroscopic scan of the area of interest. Iodine contrast is then injected via a catheter and a second 'fill' scan is undergone. The two scans are identical, except the mask does not contain any contrast. As they are otherwise identical, the mask scan can be logarithmically subtracted from the fill scan to remove unnecessary background structures:
lnIs = lnIc - lnIm
where Is is the subtracted image, Ic is the contrast-filled image and Im is the mask image. The reason logarithmic subtraction is used instead of normal subtraction is because x-ray attenuation is approximately exponential and normal subtraction would result in artefacts.
Road mapping
Road mapping is a specific technique used in DSA, which builds on the last frame hold technique in digital fluoroscopy. A DSA sequence is performed and the frame with maximum vessel opacification is obtained. This image becomes the 'road map mask' from which subsequent images is subtracted, producing a real time fluoroscopic sequence overlaid on a static image of blood vessels. Road mapping is most commonly employed in catheter insertion and guiding wires for aneurysm coiling.
Dual Energy Subtraction
Dual energy subtraction is a method used to reduce anatomical noise. DES exploits the difference between the atomic numbers of bone and soft tissue by taking two images using either a single detector and two exposures, or two detectors (sandwich) and a single exposure. The two images are then subject to logarithmic subtraction, which is weighted by R. R can be manipulated to enhance certain structures, for example, bone in one case but soft tissue in another case.
lnI(high) - RlnI(low)
For a single detector, two images are taken, one at high and one at low kVp, in quick succession. For a sandwich detector, a single exposure is made. The lower energy photons are absorbed by the front detector and the higher energy photons are absorbed by the back detector.
Spatial Resolution
Spatial resolution is the size of the smallest object than can be resolved by the imaging system. High spatial resolution improves image quality. Spatial resolution can be tested using a phantom which a series of lines which indicate from low to high resolution. Spatial resolution is affected by:
- System detectors
- Focal spot size
- Geometry
- Motion
- Pulsed fluoroscopy
Contrast Resolution
Contrast resolution is the ability to detect changes in the density of different structures in the image. It is measured using a phantom with discs of varying thickness. Contrast resolution in fluoroscopy is generally poor compared to diagnostic radiography. Contrast resolution is affected by:
- kVp
- mAs
- Scattered x-rays
- Veiling glare in image intensifiers
- Filtration
- Collimation
- Imaging processing
- Contrast media
Image Quality
Image quality is a measure of the number of details detected and number of bar patterns resolved in the image.
Comments
Post a Comment