Exposures I Chapter 20

digital radiography

replaced film with reusable detector

CR (computed radiography)

images are acquired on imaging plates that must be processed by a CR reader

DR (digital radiography)

images are acquired on a FPD that communicates with a computer for image processing

• direct conversion

• indirect conversion

digital radiographic systems are divided into two classes:

indirect conversion

convert x-rays to light, light to electronic signal

direct conversion

convert x-rays to the electronic signal

analog information

can assume infinite number of values

• continuous/smooth

• no EI/DI valus

digital information

can only assume a few discrete values

• limited by digital display made by pixels

• assigned by numbers (0 or 1)

binary system

digital language of the computer

• 0 = current OFF = black

• 1 = current ON = white

0

current OFF = black

1

current ON = white

binary digit (bit)

each 0 or 1

byte

group of 8 bits (combination of 0s or 1s)

• each byte represents shades of gray

matrix

square series of boxes that form the image

pixels

each box within the matrix

• picture element

voxel

each pixel represents 3D volume tissue of patient.

field of view (FOV)

size of the whole image

Does the FOV change?

will always remain the same = CONSTANT

spatial resolution

geometric sharpness of the image

• important in mammography

matrix size

what controls the spatial resolution?

increase number of pixels = decrease pixel size = BETTER spatial resolution

when matrix size increases, what happens to the number of pixels and the pixel size, and the spatial resolution?

• Largest matrix

• smallest pixel size

• smallest pixel pitch

• increased scanning frequency

How do you achieve the best spatial resolution?

grayscale bit depth

The number of bits used to represent each pixel

• The range of shades a pixel can show

• bit depth

• pixel depth

Other names for grayscale bit depth

2ⁿ

How do you calculate for the number of possibilities of shades of gray?

Contrast/density resolution

refers to the number of shades of gray

Bit depth

What controls contrast/density resolution?

more possibilities of shades of gray = increases contrast resolution = BETTER spatial resolution

When bit depth is increased, what happens to the contrast resolution and spatial resolution?

spatial location domain

refers to the data in the form of matrix, with each pixel assigned bits that represent a specific shade of ray.

• image is stored as a matrix

• each pixel has number that tells shades of gray

spatial frequency domain

refers to the data in the form of frequencies

• how often the image changes from light to dark

frequency

number of times a signal changes per unit length

• Increases contrast (has lots of signal changes = black and white)

• makes the image sharp and detailed

• for small, bony details

what happens to the contrast when you have a high frequency and where does this occur?

• decreases contrast (doesn’t have much signal change = gray and darker gray)

• smooth/blurry image

• for large objects with less details

what happens to the contrast when you have a low frequency and where does this occur?

Fourier transformation

used to change data from the spatial location domain to spatial frequency domain

to help the computer analyze image and improve quality

What is the purpose of Fourier Transformation?

Point-processing operations

Software functions that require the data in spatial location domain, has to be in matrix.

• happens after image is taken and before it’s shown on the screen

• rescalling

• histogram

• LUT

What involves in point-processing operations?

rescalling

software function that changes image BRIGHTNESS

• makes up for OVEREXPOSURE or UNDEREXPOSURE

Makes it lighter

How does the computer improve overexposed images through rescalling?

Makes it darker

How does the computer improve underexposed images through rescalling?

histogram

A graph of shade of gray versus the number of pixels with each shade of gray

• compares captured image with perfect histogram

wide histogram = increases contrast (both black and white pixels present)

How does a wide histogram affect contrast?

narrow histogram = low contrast (pixels centered around shades of gray)

How does a narrow histogram affect contrast?

Look-up-table (LUT)

Mapping function in which all pixels are changes to a new gray value

• adjusts CONTRAST

• Computer has LUT for each anatomy stored in memory

• apply a predetermined scale of contrast to the digital data set

windowing

Adjustments by technologist

window level

Technologist adjusting for BRIGHTNESS

window width

Technologist adjusting for CONTRAST

wide window width = decreases contrast (lots of shades of gray with few difference in between)

What happens to the contrast with a wide window width?

narrow window width = increases contrast (few shades of gray with big difference in between)

What happens to the contrast with a narrow window width?

global processing operation

uses ALL pixel values from old image to create 1 pixel value in the output image

• must be in spatial frequency domain

high-pass filtering

passes high frequency data, eliminates low frequency data

• sharpens the image

• HIGH contrast

low-pass filtering

passes low frequency data, eliminates high frequency data

• smoothes image/noise reduction

• LOW contrast

geometric processing operation

allows the tech to adjust for rotation and magnification and flipping the image.

Low contrast resolution

the ability to see few difference of shades of gray

digital system

What system has a better low contrast resolution (easier to see low contrast)

noise

unwanted random signals from the image

signal to noise ratio

amount of signal compared to amount of noise

mAs

what controls the signal to SNR?

• x-ray photons striking the IR

underexposure

What causes noise?

high SNR = good resolution (more signal than noise)

What happens to resolution with high SNR?

low SNR = noisy image (more noise than signal)

What happens to resolution with low SNR?

Increased noise = decreases contrast

When noise is increased, what happens to the contrast?

Increase mAs

How to solve for a low SNR?

Quantum mottle

caused by underexposure

• too few photons striking IR

• Can result on image missing information

• occurs on both DR + CR

Detective quantum efficiency (DQE)

Comparison between photons that strike the detector and photons that contribute to useful image

DQE = 1

what number is the perfect DQE?

No

• DQE = ALWAYS lower than 1

Will you achieve a 1 DQE?

• some photons are negated by noise

Exposure indicators

amount of exposure IR has received

• represent exposure to a digital detector

• may have a direct or inverse relationship with exposure

• can be used to assess image quality

Fuji

uses sensitivity number or S number

• has INVERSE relationship with IR exposure

Kodak (carestream)

Uses EI

• Has direct relationship with IR exposure

Agfa

Uses LgM

• Has direct relationship with IR exposure

+3 to -3

what is the range of DI?