Advanced physics unit 3

Preprocessing

-All corrections made to "raw" data due to physical flaws in image acquisition (acquisition processing)

Post processing

-All manipulation and adjustments made after corrections have been made for image acquisition (refinement)

EFR

-Exposure Field Recognition -The computer must distinguish btwn: dark structures within the anatomy from pitch black background density, which must be eliminated from calculations.

Noise reduction for DEL drop out

-Individual DELs can drop-out due to TFT failure or other malfunctions -leads to loss in information -Kernel

PPR

-Partial Pattern Recognition -Accumulation of too much background and scatter radiation btwn fields may cause software to mistake them as anatomy within the field -Result: each exposure is processed to improper level

Field uniformity

-Several flaws in the electronic and optics of the receptor and processor must be corrected

Kernel

-Sub matrix used to compensate for missing data from the del drop-out -kernel information is averaged out and used for the missing DEL data -Overtime DELs can wear out and may need plate replacement

Flat field uniformity

-evening out the overall brightness or signal across the entire imaging plate.

rescaling

-process that makes image appear normal (adjusted brightness) -Algebraically applying he same labels to incoming data regardless of what the actual data is -Begins with a pre set LUT

4 steps of preprocessing

1. Field uniformity corrections 2. Noise and Del drop-out corrections 3. Image and histogram analysis 4. Rescaling (processing)

List the 3 general approaches to processing any digital image

1. Spatial domain 2. Intensity Domain 3. Frequency Domain

any image layer can be

1. boosted (amplified) 2. suppressed (de-ampliphied) 3. removed from stack upon recomposition

2 things the computer can do..

1. can align the average brightness of the image 2. can rescale the gray scale

4 parts of image analysis

1. segmentation - CR only 2. exposure field recognition 3. initial histogram construction 4. histogram analysis

image analysis

Before any image manipulation can occur, analysis of the acquired data must be done

sub-matrix averaging to correct for DEL drop-out

Kernel

For gradation processing, the rescaled data set is fed into a anatomical ... that was determined when the operator selected the procedure at the console.

LUT

list of Q values assigned to each exam

LUT

Shifting the position of the gradient curve to the left or right has the effect of aligning the entire image histogram to the "ideal" reference ... for that anatomy

LUT/histogram

step 2 of preprocessing

Noise reduction for DEL drop-out

Simply counting the number of pixel w/in he image at each density level. Displayed as a bar graph.

Preprocessing: step 3 image analysis, initial histogram construction

Used in both CR and DR. Identifies the pitch-black densities. Recognizes landmark changes in the histogram. VOI.

Preprocessing: step 3, EFR (exposure field index)

Used in CR to identify and count the number of exposures on the plat. PPR software identifies the edges of the exposure field. Histogram errors can occur from 2 or more exposure fields being counted as one.

Preprocessing: step 3, segmentation

step 4 of preprocessing

Rescaling

area of information relevant to image formation

VOI

the gray level of each pixel becomes the ... of the waves in this graph

amplitude

the computer can not do

can not alter the pixel count in each bin

histogram analysis errors can be effected by...

collimation & scatter

Preprocessing makes...

corrections

remnant current flowing in DR

dark current

background exposure left on CR PSP

dark noise

the separated frequencies can each be amplified, modified, or eliminated individually, before ... them into the final image.

decomposing

high frequency means that there are fewer pixels per detail in the image, and this translates into higher image ...

detail

... is the range of different gray levels made available by a computer system including its installed software.

dynamic range

the net result is .... (high-pass filtering)

edge enhancement

unsharp mask filtering

edge enhancement

an example of different terms used by different manufacturers is GE's "look" feature: its "hard" look setting actually angles ...., and the "soft" look setting actually engages a ... algorithm

edge enhancement, smoothing

compression of the gray scale

equalization

corrections including flat-field, gain offsets, light guide variation

field uniformity

the mathematical method by which a complex wave form representing one line of a digital image can be broken down into a set of the different individual wavelengths that make up is know as ...

fourier transformation

domain that sorts according to size of object or details

frequency

detail processing can be performed either in the ... domain or in the ... domain

frequency, spacial

gradual passing from one tint to another

gradation

the thicker the phosphor scintillator layer, the ... absorption efficiency

greater, produces more light in the response

to achieve exposure indicators within the correct recommended range, it is essential that .... kVp, .... mAs technique combination be utilized

high, lowest

Sorting an image by the intensities of its pixels results in a ....

histogram

bar graph representing pixel count vs pixel density

histogram

step 3 of preprocessing

image analysis

A look up table really is a simple table with just two columns: one for ... and one for ...

input, output

domain that is sorted strictly by that values of the pixel

intensity

the larger the size of the ... matrix used, the wider the region use for averaging, and the more "blurred" the mask image appears

kernel

high pass frequency filtering does not replace some of the ... frequency layers (gross structures)

low

in an image that already has ... contrast, applying smoothing can lead to loss of detail.

low

to resize an image, the pixels are mapped onto a smaller or larger image ...

matrix

Nearly all modern CR and DR systems can be operated at a speed class of 300, 350 or 400 without the appearance of substantial .... in the image.

mottle

insufficient techniques resulting in low exposure indicators can cause an unacceptable level of ... noise in the image

mottle

smoothing algorithms, sometimes called "noise compensation," can be used to suppress moderate amounts of .. in an image

mottle

Updates histograms averaging data from previous 50 procedures for specific anatomy

neural histogram

final preparation for display of the image includes at least three steps 1. ... reduction 2. additional ... processing 3. ... for display

noise, radiation, formatting

two general types of noise are ... noise and ... noise

periodic, random

when PACS images are brought up for display on a particular monitor screen, sizing of the ... must fit the monitor, and the scaling of brightness and contrast must match the dynamic range of the display monitor.

pixel

computer system cannot adjust this value

pixel count

aka image refinement

post processing

corrections for flaws in the original image acquisition

pre processing

Fire-Optic guides are used to direct the light emitted to the PM. Differing lengths, quality, and flaws of the optics can vary the light intensity. Software corrections can compensate for the non-uniformity.

preprocessing: Field uniformity, light guide variations in CR

histogram is never changed from original

priori histogram

Post processing makes...

refinements

radiographs presenting severe mottle from underexposure should be ...

repeated

normalizing, brightness adjustment

rescaling

aka spatial pattern recognition

segmentation

.... is a form of low-pass filtering that softens edges and reduces noise.

smoothing

reduction of noise on image or softening of detail

smoothing

domain that is sorted by pixel location

spacial

Kernel operations pass a smaller ... of core values over each pixel row, then column by column, assigning new pixel values.

submatrix

pitch black peak representing direct exposures to IP

tail

3 types of histograms

type 1: "tail" is present, data must not be loaded in a LUT Type 2: assumption of no "tail", large body cavity fill image receptor. Type 3: contrast agent is present, or lead shield, located both SMin &Smax to eliminate areas beyond this values.

the use of a smoothing algorithm cannot recover information lost from ....

underexposure