RIU 320 - CH 4 Instruments: Imaging Anatomy and Principle 1 (Part 2): Imaging Processor and Display
- scan conversion - Preprocessing - Persistence - panoramic imaging - spatial compounding - three-dimensional acquisition - storing image frames - cine loops - Postprocessing - gray scale - color scale - Volume (3D) Presentation - Digital-to-Analog Converter
functions of image processor
- Flat panel display has a 2D pixel matrix of thousands of LCD(or LED) elements --Elements can display different amounts of light - RBG elements: primary additive colors that present a gray scale or color display -- Red, Green, Blue -- Each element has a 24 bit color = 16,777,216 possible colors/pixel
flat-panel display elements: how many and what kind of elements? what can they display? other kind of elements? bit color of each element?
the signal will go down a row and go back up; every shade of gray in relation to location and return time will be associated with each one of the squares in the matrix
for a linear pathway, how does the signal travel
results in a series of echoes that must be located in multiple rows and columns to demonstrate reflectors
for a non linear pathway, how does signal travel
heart; shows the movements over time
what is m mode used for a lot
m-mode
what mode is this
a-mode; Probe is on left hand side of image and depth goes deeper moving to the right
what mode is this? where is probe
Most systems have 6-, 7-, and 8- bit memories
what size bit memory to most us machines use
-There must be motion of the structure and it needs to be in up and down motions in order to be displayed on m mode
what type of motion is needed for m mode
straight line
what would m mode look like for a non moving structure
not needed in rectangular scans if there is an emitted ultrasound pulse for each vertical scan line displayed
when is pixel interpolation not needed
write zoom, better resolution
which zoom is preferred
brightness increases with increases echo intensity (gray level stored in mem)
white echo display
8
1 byte = ? bits
- four-binary-digit memory would have 4 "layers" of matrix to create 4 possible bits per pixel - Bit depth = 4 - Shade memory = 2^4 = 16 (0-15 possible shades of gray)
A four-binary-digit memory would have how many layers of matrices? what is the bit depth? what is the shade memory?
- Aka Amplitude Mode: Graphical demo of signal - Y axis = amplitude - X axis = depth - One of the original forms of US - Uncommonly used (ophthalmology)
A mode: other name? x and y axes? is it used a lot?
2^8 = 256
An 8 bit system has how many possible shades of gray? (shade memory)
- The human eye can distinguish more color tints than grayscale shades, color offers improved contrast resolution capability - ability to present different echo intensities in various colors rather than in gray shades, that is, the ability to colorize echoes
B-Color: what is this and why?
Converts the digital data (numbers) received from the image memory to analog voltages that determine the brightness of the echoes on the display
Digital-to-Analog Converter
non-linear
Display scan format allows display for ____ structures through image processing
Sharpens boundaries to make them more detectable and measurements more precise - preprocessing
Edge enhancement : def? post or preprocessing?
- a backlighted liquid crystal display (LCD) or matrix of light emitting diodes (LEDs) - Computers display in horizontal lines - Each horizontal line is coded from a row of digital data from the image memory
Flat-Panel Display: what are the components? orientation of lines? how are lines coded?
a single image of one complete scan (all the scan lines)
Frame
several shades of gray or brightness and black and white
Gray scale
- Can be in white echo or black echo form (we usually use white echo) - Can be set to certain portions of the brightness range - Typically brightness increase the corresponding gray level
Grayscale Maps: two forms? how can it be set? how does brightness and gray level relate?
•Brightness (gray scale) is proportional to the echo strength •Frame: each individual image •Frame rate: number of images entered into memory per second •Refresh rate: number of times images are retrieved from the memory and presented on display per second
Image Display: brightness is proportional to what? frame? frame rate? refresh rate?
- A mode - B mode - M mode
Image Display: how can data be presented ( modes)
- Image frames are stored in the memory after echo data is preprocessed - Holding and displaying one frame out of a sequence is called freeze - Storing the last several frames acquired before freezing is called cine loop - Image memory divides the image plane into pixels
Image Memory: what does it do? freeze vs cine loops? how does it divide the image?
- Coverts the digitized, filtered, detected, and compressed serial scan-line data into images that are stored in the image memory - the direction of each scan line and the location of echoes in depth down the scan line are used to determine the proper location in the image memory (thus the display) for each echo
Image Processor general function: converts what to what? what two things are used to determine the proper location in the image memory for each echo?
1024 bytes (8192) bits
Kilobyte = ? bytes and bits
- Aka Motion Mode - Displays anatomic location over time - X-axis = time - Y-axis = depth - Pattern down a single scan line (can move the scan line)
M mode: other name? what does it display? x and y axes? what is pattern from?
- Reduces noise and smooths the image by frame averaging - takes many frames and averages them together to create a final frame: because noise is random is doesn't show up in the same pixel every time but rather an actual signal does. with averaging the actual signals will appear brighter and thus reduce the noise - Higher levels are appropriate for slow-moving structures where as low levels or off should be used for fast moving structures -Must be applied while scanning live: preprocessing
Persistence: what does it do? how? what level should be used for slow moving structures? fast? when is it applied? post or preprocessing?
- Fills in missing pixels - Assigns brightness based on average brightness of adjacent pixels - not operator controlled - preprocessing - Useful for sector scans: (because scan lines all come out of the probe at difference angles very close to each other, when they reach far depths they can be more spread out. there can be an entire pixel between the signal lines with no information. so as the machine is receiving the return signals and assigns them to their corresponding pixel location there could be holes between data. pixel interpolation helps to fill these holes by averaging the brightness of surrounding signals)
Pixel interpolation: what is it? what does it do? Why is this needed for sector scans? is it operator controlled? post or preprocessing?
- Image processing performed after the echoes are stored in memory; includes all the procedures performed with echoes while they are brought out of the memory to be displayed - assignment of specific display brightness to numbers retrieved from the memory - Operator controlled - Examples -- Gray-scale maps -- B- color -- Three-dimensional presentation
Postprocessing: when is this performed? is it operator controlled? exs?
- Image processing done before storing echo data in image memory ex: - Edge enhancement - Pixel interpolation - Persistence - Three-dimensional acquisition - spatial compounding - panoramic
Preprocessing: def and ex
white
RGB mixed equally makes what
rapid sequence of frames stored in the image memory and presented on the display
Real-Time Display
- Translates the information from vertical scan lines to horizontal video format - reformats echo data into image form for image processing, storage, and display
Scan Converter
- Acquiring several two-dimensional scans for three-dimensional volume of information in the image memory' - Aka 4D with time as the 4th dimension - preprocessing - You need to have a two dimensional array (3d probe) in order to form a 3d image: multiple layers of elements to compile 3d volume
Three-dimensional acquisition: how does it acquire the 3d image? what is the 4th dimensions technically? post or preprocessing? what do you need to take this image?
- common ways of presenting 3D echo data (volume imaging) include surface renderings, 2D slices through 3D volume (like ct), transparent views
Volume (3D) Presentation
makes it possible to present image-plane orientations that are impossible to obtain with conventional 2D scanning
advantage of 2D slice presentation
Voltage echo - analog signal - digital signal - back to analog signal to then display the image so that we can have different strengths assigned to different areas of the screen
all the conversions of the signal
b scan, brightness mode, gray-scale sonography
b mode other names
binary digital, 0 or 1
bit
bit depth = # of matrix layers
bit depth = ?
brightness increases with decreasing echo intensity
black echo display
limited by spatial pulse length (axial res) and beam width (lat res) rather than pixel density in the memory
detail resolution is usually limited by what
The display must be able to properly place that caliper and associate that with a certain distance so there is technology built into the flat panel display to be able to do this
image display: calipers
1- the beam is scanned through the patient in such a way that it cuts like a knife through the tissue cross section 2- echoes received from all points on this cross section are converted to number that are stored at corresponding pixel locations in the digital memory 3- all the information necessary for displaying this cross sectional image now is stored in the memory 4- the info than can be taken out of the memory and presented on a 2d display in such a way that the number coming out of the memory are displayed as corresponding pixel brightness on the face of the display
in summary, the procedure for entering the echo information require for display of the 2d cross section image into digital memory is as follows:
various instruments may store echo info at various points along the processing chain, includes storing radio frequency echo data before amplitude detection
memory of info can include more than just the image memory: what else can be stored
improves dynamic range and contrast resolution
persistence: speckle (noise) reduction improves what
- frame averaging reduces frame rate because averaged frames are no longer independent - operator control permits averaging of a selectable number of sequential frames
persistence: what is the effect of frame averaging on frame rate? operator controlled?
binary, 0 or 1, one or off
pixel format (memory element) in each matrix is what
- squares that divide the 2D image plane into a matrix - Typically several thousand/frame - a number the corresponds to the echo strength received from the location of anatomy corresponding to that memory position is stored in each of the pixel locations in the memory
pixel: what are they? how many? what is stored in pixel locations?
- Must have more than one matrix for gray scale - uses binary bits
pixels and bits: how do we get more than just black and white in a single matrix?
- each pixel could only store one of two numbers: 0 or 1 - this would mean it only shows black or gray (bistable), but for US we need more grays - binary
pixels: single layer matrix : how do pixels store the info?
the type of magnification performed in the D-to-A converter (postprocessing) that magnifies the image by enlarging the pixels; worse detail resolution
read zoom: what is it, formed in what converter, and spatial resolution quality?
- how many possible shades of gray can be assigned to a frame - equation = 2 ^ bit depth
shade memory: what is it? equation
- digital: stored as numbers (binary; digits)
sonographic image memories are what format?
from signal processor -- preprocessor -- scan converter -- image memory -- post processor -- digital to analog converted -- to display
steps in image processing
obstetric and echocardiography
surface rendering are popular in what imaging
allow see through imaging of anatomy, as with plain film radiographs
transparent views
-Low persistence: more noise -High persistence: less noise
high persistence, ___noise? low persistence ____noise
the can be equal, but not always
how are frame rate and refresh rate related
each gray scale pixel is represented in eight binary digits, 8 bits
how is each gray scale pixel represented
Humans can see about 100 shades
how many shades of gray can human eyes see?
determines the depth to be displayed and thus the depth range covered by the image memory in the scan conversion process
image depth control determines what
- Can be 2D or 3D - Displays spots of brightness at specific locations, usually in relation to anatomy - what we usually use
image display: B mode: dimensions? what does ti display?
Have to be careful about using persistence with moving structures because they will appear blurred because the object appears in different pixels/frames each time, so when all the frames are averaged together is appear blurred
why is persistence no good for fasting moving structures?
because we want to display pixels with proportional brightnesses, not the numbers that represent the echo intensities in the image memory, these number must be converted by the DAC to proportional voltages that control the brightness
why is the digital to analog converted needed to see things
the type of magnification performed in the A-to-D converter (preprocessing) that magnifies the image by redrawing it (reassigning pixels) before it is stored in memory; increased spatial resolution
write zoom: what is it, what converter and spatial resolution quality?