| |
| Editor's Note |
| |
| The first four pages of
the instrumentation section is a technically detailed discussion of
the acquisition, processing, output of the captured images, and compression
formats of those images. If you do not require detailed knowledge
of these topics, please skim over the sections and proceed to the knobology section. While the initial information on instrumentation
is academically interesting, its contribution to clinical echocardiography is more limited than later sections. |
| |
| Objectives |
| |
| At the completion of this chapter, the student will be able to: |
| |
- Discuss Digital Image Compression
- Discuss Basic Knobology
- Discuss 2D Controls
- Discuss B Mode Controls
- Discuss M Mode Controls
- Discuss Color Controls
- Discuss Pulse Wave Doppler Controls
- Discuss Continuous Wave Doppler Controls
|
|
| |
| Image Generation and Processing |
| |
| Digital
vs. Analog |
| |
| Analog |
| |
| Modern
echocardiography primarily performed in a digital fashion. Digitalization
of the ultrasound signal offers many advantages over the analog signal.
Image 3.1 shows a graphic view of the capturing, manipulation and
display of an ultrasound signal. While the original transmission (1)
and reception of the ultrasound signal (2) is analog, modern
echocardiographic machines convert the received signal to a digital
signal(3). After digitization of the signal is completed, various
processing (4) of the digital signal is performed. After the
processing is completed, the digital signal is either saved (to a
disk and/or in computer memory) or it is converted back to an analog
signal (5) for display to a monitor screen where it can be
recorded to a videocassette tape. This analog-to-digital-to-analog
conversion is common to most echocardiographic machines. If the machine is totally digital, the analog conversion is not done and the digital processing results in a digital image/video that is displayed on a digital monitor. |
| |
|
|
|
|
|
|
|
|
1-Transmission |
2-Reception |
3-Digitalization |
4-Processing |
5-Analog Conversion |
6-Display |
|
|
|
Image 3.1 |
| |
| Digital |
| |
The most modern echocardiographic
machines perform an analog-to-digital conversion whereby the
information is stored, displayed, and processed for display from
its digital format. The digital format will not degrade when multiple
copies are made, it is electronically searchable, it can be transmitted
over a network, provides a superior image quality, and it can be
processed after the signal is recorded.
|
| |
| Digital and Analog Comparison |
| |
| Factor |
Digital |
Analog |
| Image Quality |
Excellent |
Fair |
| Processing after recorded |
Yes |
No |
| Transmission |
Network |
Mail |
| Copy Quality |
Same |
Degrades |
| Searching |
Random Access |
Sequential |
| |
|
Table 3.1.2 |
|
| |
| Binary Representation
of Digital Information |
| |
| Digital information, whether it is images, text,
or sound are stored in a series of ones and zeros on digital media.
The placeholder for the zero or the one is called a bit. Depending
upon the computer system used, the bits are stored in groups called
bytes. If a byte is a grouping of 4 bits, then it can store four
zero's or four one's or any combination of one's and zero's as long
as it is 4 bits long.
In the binary world, numbers are derived from bytes. In the table
below, as we change the bits from zero's to ones, the numbers from
0 to 15 can be represented in a single byte. The bits, since they
can only be a zero or a one are a base two representation.
Each bit is raised to the power or 2. The first bit on the right
is 2 to the power of 0, the second bit on the from the right represents
a 2 to the power of 1, the third bit represents a 2 to the power
of 2, and the third bit, the one on the left, represents a 2 to
the power of 3.
For example, if a byte had the four bits equal to 0100, then the
number represented would be 4 because the third bit is 22 or 4. The table shows how each number from 0 to 15 are represented
in the binary world. |
| |
|
Byte |
Calculation |
Number |
|
0000 |
0 * 23 + 0 * 22 + 0 * 21 + 0 * 20 |
0 |
|
0001 |
0 * 23 + 0 * 22 + 0 * 21 + 1 * 20 |
1 |
|
0010 |
0 * 23 + 0 * 22 + 1 * 21 + 0 * 20 |
2 |
|
0011 |
0 * 23 + 0 * 22 + 1 * 21 + 1 * 20 |
3 |
|
0100 |
0 * 23 + 1 * 22 + 0 * 21 + 0 * 20 |
4 |
|
0101 |
0 * 23 + 1 * 22 + 0 * 21 + 1 * 20 |
5 |
|
0110 |
0 * 23 + 1 * 22 + 1 * 21 + 0 * 20 |
6 |
|
0111 |
0 * 23 + 1 * 22 + 1 * 21 + 1 * 20 |
7 |
|
1000 |
1 * 23 + 0 * 22 + 0 * 21 + 0 * 20 |
8 |
|
1001 |
1 * 23 + 0 * 22 + 0 * 21 + 1 * 20 |
9 |
|
1010 |
1 * 23 + 0 * 22 + 1 * 21 + 0 * 20 |
10 |
|
1011 |
1 * 23 + 0 * 22 + 1 * 21 + 1 * 20 |
11 |
|
1100 |
1 * 23 + 1 * 22 + 0 * 21 + 0 * 20 |
12 |
|
1101 |
1 * 23 + 1 * 22 + 0 * 21 + 1 * 20 |
13 |
|
1110 |
1 * 23 + 1 * 22 + 1 * 21 + 0 * 20 |
14 |
|
1111 |
1 * 23 + 1 * 22 + 1 * 21 + 1 * 20 |
15 |
| |
|
Table 3.1.3 |
|
| |
| From this table, a 4 bit system can store numbers
from 0 to 15 or a total of 16 numbers (24).
Modern systems can process vast amounts of data compared to the
4 bit system. Currently, most systems are 32 bit systems with 64
bit and 128 bit systems coming soon. The next chart shows as the
number of bits that can be stored in one byte increases, the amount
of data that can be processed increases also. |
| |
| Byte Sizes |
| |
|
Byte |
Calculation |
Largest Number |
|
2 bits |
22 |
4 |
|
4 bits |
24 |
16 |
|
8 bits |
28 |
256 |
|
16 bits |
216 |
65,536 |
|
24 bits |
224 |
1,6777,216 |
|
32 bits |
232 |
4,294,967,296 |
|
64 bits |
264 |
18,446,744,073,709,551,616 |
|
128 bits |
2128 |
3.4028236692093846346337460743177
e+38 |
| |
|
Table 3.1.4 |
|
| As you can see, as the number of bits increases
in each byte, the amount of data or bandwidth, that can be processed
dramatically increases. Besides computer processing bandwidth, displays
or images are described in these terms also. A 16 bit image or display,
can display 65,536 different colors. An 8 bit black and white image
or display will display 256 shades of gray. A 32 bit display can
display over 4 billion colors, which is more colors than the human
eye can recognize. |
| |
| Digital Storage |
| |
| Storage of information on media uses a similar language
to describe the capacity of the media. Storage media comes in all
types, sizes and access speed. After punch cards and magnetic tapes came the floppy
disc in a 3 1/2 inch or 5 1/4 inch size. Since then zip disks,flash drives, CDs,
DVDs, Magneto Optical Disks, hard drives, digital tape, and RAM
have all been invented or improved. Larger capacity is required
for digital information because of it's size. To display a saved
image, the access time of the media will determine the speed of
it's display. The capacity of the media is described in bytes. However,
unlike processing or display bytes, a byte of storage media represents
a 8 bit wide byte. An system where a byte is 8-bits wide can store
a character, such as a letter, a number, or a special character
(i.e. decimal point). The size or capacity of the media is
then described not by increasing bits but by the number of 8-bit
bytes that can be stored. The table below describes common capacities
of storage media: |
| |
|
Media |
Number
of Bytes |
Access
Speed |
| 5 1/4†Floppy |
720 Kilobytes |
seconds |
| 3 1/2†Floppy |
1.44 Megabytes |
seconds |
| Zip Disk |
100 or 200 Megabytes |
seconds |
| RAM |
12 Gigabytes |
nanoseconds |
| CDROM |
650 Megabytes |
seconds |
| Magneto Optical Disk |
650 Megabytes |
seconds |
| Flash Drive |
128 Gigabytes |
milliseconds |
| DVD |
5.4-9.6 Gigabytes |
seconds |
| Hard Drive |
1-4 Terabyte |
milliseconds |
| Digital Tape |
1-10 Terabyte |
minutes |
| |
|
Table 3.1.5 |
|
| |
A kilobyte
is 1024 (one thousand) bytes. A megabyte is 1,048,576 (one million)
bytes. A gigabyte is 1,073,741,824 (one billion bytes), and
a terabyte is 1,099,511,627,776 (one trillion) bytes. While
some of these capacities seem almost infinite, digital video
is also quite large. Uncompressed video will run about 622,080,000
bytes or 600 megabytes per minute (720 x 480 video at 30 frames
per second for one minute). If a study takes 5 minutes of digital
video, the file will be 3 gigabytes. Even if 10 hard drives
were daisy-chained together (a RAID system), the total capacity
of the RAID system would be 3 terabytes or about 900 studies.
Clearly, a modern echocardiography system will require fast
computers, large hard drive RAID systems, and a digital tape
for long term storage.
Newer echocardiography machines will sample video at 60 frames per second at a video size up to 1200 x 800 resulting in a video of about 9 gigabytes per study. A higher resolution study will tremendously increase the amount of storage space and network speed that is required for a modern digital echocardiography system.
|
| Storage Relationship |
| 1 Byte = 8 bits |
| 1 Kilobyte = 1024 Bytes |
| 1 Megabyte = 1024 Kilobytes |
| 1 Gigabyte = 1024 Megabytes |
| 1 Terabyte = 1024 Gigabytes |
| 1 Petabyte = 1024 Terabytes |
| 1 Exabyte = 1024 Petabyte |
| 1 Zettabyte = 1024 Exabyte |
| 1 Yottabyte = 1024 Zettabyte* |
| 1 Brontobyte = 1024 Yottabyte* |
| 1 Geobyte = 1024 Brontobyte* |
| 1 Zotzabyte = 1024 Geobyte* |
| * = Proposed Table 3.1.6 |
|
|
| |
| Processing Requirements |
| |
| Most of the current echocardiographic machines have
a single CPU. A single processor has limits to its speed of processing
the vast amount of bits of information from a video display. The
limit in an echocardiographic machine is the number of pulses that
can be sent and received in a set amount of time. This limit
is called the pulse repetition frequency. The speed of sound, c, and the depth of
the display, will determine the pulse repetition frequency (PRF).
The formula for a pulse repetition frequency is: |
| |
| PRF = c/2d |
| |
| where c = 1540 m/sec and d is the depth. At a depth of 12 cm the
PRF will be: |
| |
| PRF = (154000 cm/sec)
/ (2 * 12 cm) |
| |
| which is 6,417 cycles/sec or 6.417 kHz. Therefore, at 12 cm the
maximum PRF or pulses per scan line is 6417/sec. If a scan has 100
scan lines per frame then the total pulses per frame would be 64,170
pulses per frame per second. If the frame rate is 30 frames per
second (fps) then the number of pulses would be 2,139 per frame.
The frame rate, scan lines, and pulses per scan line are all interrelated.
If one factor is increased then one or both of the other factors
must be decreased. The relationship between pulses per scan
line, scan lines per frame, frames per second, and the PRF is shown
below: |
| |
| PRF
= pulses/scan line x scan lines/frame x frames/sec |
| |
| Resolution is the ability to discern between two
distinct anatomic points. The displayed resolution depends upon
the above factors. The number of pulses along a scan line determines
the resolution along that line. The number of scan lines per frame
determines the resolution between scan lines. The frame rate determines
the resolution of cardiac movement or events. Since all of the factors
are related to produce a PRF value. Since the PRF determines the
maximum of the other factors, manipulating the other factors determines
the quality of the image displayed. The best scan will have a high
frame rate, a high number of scan lines, and a high number of pulses
along the scan line. The frames per second could be increased but
at a cost of the number of pulses per second or the number of scan
lines per frame. The number of scan lines could be increased but
at the cost of the number of pulses or the frame rate. Lastly, the
number of pulses could be increased but at a cost of the number
of scan lines or the frame rate. |
| |
| Raster
vs. Vector Images |
| |
| Images are displayed in two basic formats: raster
or vector. Raster images are pixel based. Raster images are drawn
and manipulated by changing each pixel. A pixel is a "dot"
on the screen that represents the resolution of the image. If an
image is 720 x 480 pixels, then it will contain 720 * 480 or 345,600
pixels. Each pixel can represent a color depending upon its bit
depth. As noted above if the image is an 8 bit image, then each pixel
has the possibility of one of 256 colors. As the number of pixels
increases the resolution of the image increases. True photographic
quality is 2400 x 2400 pixels whereas digital video is 720 x 480
pixels. Most monitors will display 800 x 600 or higher resolutions.
Raster images, if enlarged or shrunken, will appear distorted from
the original image. As the image size is changed, the pixels between
other pixels must be filled in or deleted. This filling in or deletion
phenomena will cause the image to become blurry and distorted because
the program must guess the pixel values between the original pixels.
Below is a raster image. A portion of the raster image has been blown
up in the red circle. Inside the red circle is blown up further
and displayed in the enlarged raster image picture. From a distance
boundaries will appear sharp, but, under enlarged conditions, the
boundaries are blurred as the pixel values change from white to black.
When the image is shrunk, boundaries also become blurred because
pixel values must be filled in where no original pixel value existed. |
| |
 |
 |
 |
Image 3.1.7
Normal Raster Image |
Image 3.1.8
Enlarged Raster Image |
Image 3.1.9
Reduced Raster Image |
|
| |
| Vector images, on the other hand, do not suffer
distortion when their size is changed. Vector images draw the image
using a series of mathematical algorithms. If the size of the image
is changed, the image is simply redrawn using the mathematical formulas.
Vector images always appear sharp at any size. However, Vector images
cannot draw each pixel differently from adjacent pixels. All of
the pixels in a given area are the result of a mathematical formula.
Therefore, for pixel-to-pixel control raster images are superior,
but, for image size manipulation, vector images are superior. |
| |
 |
 |
 |
Image 3.1.10
Normal Vector Image |
Image 3.1.11
Enlarged Vector Image |
Image 3.1.12
Reduced Vector Image |
|
| |
| Voxels |
| |
| While an image or video is displayed on a 2D screen, it is displayed in
pixels. However, 2D images and video are captured in pixels, whereas, 3D
volumes are captured in voxels. A voxel is a unit of volume, similar to
a pixel as a unit of area. Voxels are combined to result in a volume
set. Pixels are combined to result in an image set. Voxels are raster,
not vector based so similar rules apply when altering a voxel based
video. |
Voxels |
|
| |
| |
| |
