Digital Image Acquisition and Processing

Radiographic Exposure

bit

binary number (0 to 1)

byte

8 bits

Digital radiography

uses reusable electronic detectors that acquire attenuation information and display it as binary numbers on a computer where it can be manipulated

Computed Radiography (CR)

a digital imaging process that uses a PHOTOSTIMULABLE PLATE to produce images

What is CR composed of?

europium-doped barium fluorohalide crystals

CR PSP layers

1. protective layer

2. phosphor layer - active layer with barium fluorohalide crystals

3. support layer -coated with active crystals

4. conductive layer - increases sharpness by absorbing light

CR is approximately equal to a _____ combination.

200 speed F/S

CR phosphors give off _____ through _____ absorption of x-ray.

light

photoelectric

DQE (detective quantum efficiency)

the input squared to output squared signal from the IR

K-shell absorption edge means absorption through _____ processes producing _____.

photoelectric

characteristic radiation

Latent image formation (2 things):

1. x-ray hitting plate causes photoelectric interactions

2. photoelectrons EXCITE electrons

The latent image loses _____ of its energy after 8 hours.

25%

The “F” Trap

a location in the CR phosphor crystal that holds elevated electrons

The “F” Trap: X-ray causes electrons to elevate from _____ energy band to _____ energy band.

valence

conductive

Electronic flat panel detectors contain

DEL’s (detector elements)

REMEMBER: DEL size has the most dramatic effect on resolution of any variable

Indirect Digital Flat Panel Components use two things:

1. scintillator

2. photodiode

What does a scintillator do?

convert x-ray to light

What does a photodiode do?

converts light to an electrical signal

What is a scintillator composed of?

rare earth phosphors and cesium iodide (CsI) crystals

Photodiode is composed of what?

Amorphous silicon (converts light to electrons)

Thin Film Transistor (TFT) (4 steps):

1. x-ray strikes the scintillator emitting light

2. light is picked up by photodetector (photodiode) and changed to an analog electrical signal

3. the electrical signal is picked up by a flat panel detector and sent to the ADC (analog-to-digital) of the computer for active readout

4. the analog electrical signal to histogram

A charged couple device is an indirect system composed of what?

tiled silicon chips

Process of charged couple devices (3 steps):

1. fiber optic cables focus light from the scintillator on the CCD chips where electron holes are created in the silicon representing the image

2. the electron holes are held as electronic charges

3. the charges are read out row-by-row by applying small voltages to reach DEL and sending them via TFT’s

What do Complimentary Metal Oxide Semiconductors (CMOS) IR’s do (5 things)?

1. uses a scintillator

2. light is captured by capacitors

3. transistors used to switch and amplify the charge

4. light photons converted to electrical charges

5. an ADC turns pixel’s value into digital value

LOOK AT SLIDE 32

Photoconductor and what does it consist of?

changes x-ray to analog signal

amorphous selenium

Detector elements with thin-film transistors

act as storage area and electronic gates to transfer electrical signals for processing

Fill factor

the percentage of the DEL dedicated to photon absorption

Direct capture process (4 steps):

1. prior to exposure a high voltage is applied to the top of the selenium plate

2. radiation striking the selenium causes the release of electrons in relation to the exposure received

3. the electrons are captured by charge collection electrodes at the base of the selenium

4. the charge is located by the TFT and sent

Power of 2 is used to designate what three things?

1. image size

2. dynamic range (shades of grey)

3. storage capacity

CR image processing: the _____ converts the analog latent image to a digital manifest image

image plate reader

LOOK AT SLIDE 46

Placement of collimated fields of CR should include

4 collimated edges (3 at min.)

Sampling frequency determines the _____ of digital images & _____.

contrast resolution

spatial resolution

Nyquist criterion says digital images must be sampled _____ for each analog voltage cycle.

twice

Aliasing

improper sampling caused by the spatial frequency exceeding the Nyquist frequency

The maximum sampling frequency depends on the _____ of CR or the _____ of digital flat panel detectors.

sampling pitch

DEL pitch

Fixed sampling

maintains a fixed sampling frequency to maintain fixed spatial resolution regardless of IR size

Fixed matrix size

spatial resolution changes based on the size of the IR

The histogram is representative of _____ and _____ for an exam.

specific anatomy

technique

Quantizationon

process of assigning a digital number to each pixel to represent brightness levels in the image

The number of brightness levels represents various _____ values of anatomy. The _____ the number of shades of gray that can be displayed by the computer, the _____ the gray-scale bit depth.

attenuation

greater

greater

Digital systems can display as many as

2^14 (16,384) bit depth

The histogram shows _____ on the far left with soft tissues and ______ on the right with raw radiation on the extreme far-right spike.

bone

aerated lung

The far right shows

Pre-processing and 3 examples

occurs automatically prior to original image being displayed

• field uniformity corrections

• histogram analysis

• rescaling of image

Post-processing and 3 examples

occurs after image is displayed

• masking

• inverting

• magnification

Fourier transform changes information in the _____ and converts it to the _____.

spatial location

frequency location

Point processing operations is performing operations on each _____ based on _____ location.

individual pixel

pixel

Grayscale processing is one example of _____ where a histogram is created and compared to a look-up table in order to adjust the image brightness and contrast

SLIDE 70 & 71

point processing

Local Area Processing is the application of _____ that are applied to small _____ of pixels.

kernals

groups

High-pass filtering (edge enhancement):

Edge enhancement whereby low spatial frequencies are removed by Fourier transform to produce sharpened or enhanced edges

Low-pass (smoothing):

filtering removes excessive image noise  and
intentionally blurs the image (may result in some loss of image detail)

Un-sharp masking

Subtracts a low-pass filtered image from the original image to produce a new subtracted and sharper image

Spatial location filtering (convolution)

Applies a kernel across all pixels
to apply a coefficient across the entire matrix. Can be used to hide dead pixels

Geometric processing

A type of processing that changes the position or
orientation of pixels in the entire image - Examples include rotation, magnification,
flipping (mirroring)

What is a kernel?

a submatrix that is passed over a larger matrix that executes some mathematical function

The kernel then indexes _____ to _____, repeating the calculations until the _____ image is covered.

left to right

entire

High pass filtering (edge enhancement)

a filtering technique that improves image contrast by bringing out high frequencies in the image

Over-use of edge enhancement can result in an increase in _____ and _____.

image noise

“halo-effect artifacts”

Low-pass filtering (smoothing)

a technique to remove noise from the image

Smoothing Kernel

represented by kernel A where all values hold the same value

Edge-enhancement kernel

represented by kernel B where the cells sum to “1”

SLIDE 81