Creation of a digital image from an analog signal. Analog-Digital Converter (ADC)

1
Creation of a digital image from an analog
signal. Analog-Digital Converter (ADC)
2
Creation of a digital image from an analog
signal. Analog-Digital Converter (ADC)
3
Creation of a digital image from an analog
signal. Analog-Digital Converter (ADC)
4
Pixel Saturation
Saturation can be described by the fact that no
pixel can be darker than pure black (i.e. value
0) nor brighter than absolute white (i.e. greater
than 255 or 4095 or 65,535)
5
Pixel Saturation
Once a pixels value is saturated it can no
longer provide useful information
6
The data contained in a digital image can be
displayed as a histogram which is a plot of the
pixel values ranging from black to white versus
the number of pixels that have that particular
value.
Histograms
7
Low Contrast image (pixels concentrated in a
narrow, central values) High Contrast image
(many pixels clustered at the extremes of the
value range)
8
Good contrast (even distribution)
9
One can take an image with a narrow contrast
range and expand it to cover the entire range of
black to white in a process known as contrast
stretching.
10
The display values of those pixels (i.e. how they
will appear on the viewing screen) is determined
by the display curve. In the histogram to the
left a pixel value of 68 would have a display
value of 68. The display curve is straight with
a slope of 1.00
Histograms
11
The slope of the display curve relative to the
pixel values can be varied in any number of ways.
This is known as a gamma correction and can be
used to
expand the display of either the dark (g lt 1) or
bright (g gt1) portion of the image.
12
Slope 1.0 0.5
2.0
Gamma correction can be easily adjusted in many
image processing programs by choosing the
midpoint of display on the histogram of pixel
values and thus changing the slope of the display
curve.
13
micro.magnet.fsu.edu/primer/digitalimaging/index.h
tml
14
micro.magnet.fsu.edu/primer/digitalimaging/index.h
tml
15
Another way of altering the way an image is
displayed involves the use of a Look-Up Table or
LUT in which the value of the input pixel is
changed to a new value in the output. Thus a new
image matrix is created with new values for each
pixel.
16

• This affords much greater control over the
  displayed image.
• Inverted image
• All pixels below 100 are black
• All pixels between 64 and 120 are black

17
In applying a color LUT to this image pixels of
specific values are displayed as red.
18
Color LUTs or CLUT can be complex and
apparently random but if later passed through an
inverse CLUT the original image can be restored
19
Why would having additional bit-depth be
important when the human eye can only detect
about 100 shades of grey anyhow?
X-ray of Breast Contrast Enhanced
Subtle differences in grey levels contain
information
20
The fact that digital images are data matrices
means that functions can be applied to them. A
new data matrix based on the intensity of the
pixels in the background can be subtracted from
the original matrix and the resultant matrix is
then contrast stretched to produce an image in
which the distracting background has been
substracted.
21
Dodging and Burning Selectively increase or
decreasing the brightness of a subset of pixels.
22
Clone Stamp
Selectively copy a portion of the matrix onto
another portion of the image.
23
Clone Stamp
Selectively copy a portion of the matrix onto
another portion of the image.
24
Having increased bit-depth also allows one to
take advantage of sophisticated image processing.
25
Image Processing
26
Image Processing
Spatial Convolution (image processing) An image
processing operation that is used to spatially
filter an image. A convolution is defined by a
kernel that is a small matrix of fixed numbers.
The size of the kernel (3x3, 5x5, 7x7, 9x9) the
numbers within it, and a single normalizer value
define the operation that is applied to the
image. The kernel is applied to the image by
placing the kernel over the image to be convolved
and sliding it around to center it over every
pixel in the original image. At each placement
the numbers (pixel values) from the original
image are multiplied by the kernel number that is
currently aligned above it.
27
Image Processing
The operation, as defined by the kernel, is
applied to all pixels in the original data matrix
with the exception of those pixels that form the
edge of the matrix. A new output matrix of the
same size, but with different pixel values
results.
28
Image Processing
The sum of all these products is tabulated and
divided by the kernel's normalizer. This result
is placed into the new image at the position of
the kernel's center. The kernel is translated to
the next pixel position and the process repeats
until all image pixels have been processed. As
an example, a 3x3 kernel holding all 1's with a
normalizer of 9 performs a neighborhood averaging
operation. Each pixel in the new image is the
average of its 9 neighbors from the original.
29
529 / 9 59
30
(No Transcript)
31
Sharpening The sharpness of a digital image
refers to the degree of clarity in both coarse
and fine specimen detail.
32
micro.magnet.fsu.edu/primer/digitalimaging/index.h
tml
33
micro.magnet.fsu.edu/primer/digitalimaging/index.h
tml
34
micro.magnet.fsu.edu/primer/digitalimaging/index.h
tml
35
micro.magnet.fsu.edu/primer/digitalimaging/index.h
tml
36
micro.magnet.fsu.edu/primer/digitalimaging/index.h
tml
37
micro.magnet.fsu.edu/primer/digitalimaging/index.h
tml
38
Image processing can be a complex combination of
processes
First apply apply a color LUT to a greyscale
image.
39
The lowpass averaging filter can be selected from
3x3 up to 255x255 neighborhoods. All pixels
included in the filters size are added up, the
result is divided by the number of pixels.
40
The Gauss filter can be selected from 3x3 up to
7x7 kernel size. It performs a weighted sum
(center pixel highest weight). The result is
normalized by total kernel weight.
41
Process / Filters / Median With a kernel size of
3x3 the resulting value is the median (number 5
out of the sorted list of 9). In contrast to
lowpass filters, the median keeps edges and
removes single pixel errors (like hot pixels)
completely.
42
Laplace (8 connected)The Laplace Filter weights
the difference between the center pixel and its
neighbors.
-1 -1 -1 -1 8 -1 -1 -1 -1
43
Sharpen (8 connected)Based on the Laplace
Filter, the Sharpen Filter includes/adds the
original image (center pixels weight one above
Laplace)
-1 -1 -1 -1 9 -1 -1 -1 -1
44
SobelThe Sobel filter enhances edges in all
directions. It is implemented through two
independent convolutions with the left kernel
(once rotated by about 90). The results of each
kernel are combined to form the final result.
1 2 1 0 0 0 -1 -2 -1
45
PrewittThe Prewitt filter performs a similar
operation as the Sobel Filter (90 kernel
rotate).
1 1 1 0 0 0 -1 -1 -1
46
The exact sequence of processing steps can bring
about a profound change, and improvement, in the
image.
47
The spatial domain is represented by the
conventional image. The processes of adjusting
contrast, sharpening, background removal, etc.
are all forms of spatial convolution. We can
also process the data in the image matrix in its
frequency domain using algorithms developed by
Joseph Fourier
Joseph Fourier (1768-1830)
48
A Fourier Fast Transform (FFT) converts an image
into a display of frequencies displayed as an
energy spectrum. This can be done optically an an
electron diffraction pattern is an example of
this, the spacing of dots indicating their
frequency and the brightness of the dots
indicating the intensity.
49
A FFT can be created of the spatial domain and if
there is a frequency it will show up as dots.
50
A FFT can be created of the spatial domain and if
there is a frequency it will show up as dots.
If the size and brightness of the spots is
increased the frequency will be enhanced but not
changed.
51
If the processed FFT is once again put through a
Fourier transform (inverse FFT) an altered image
of the original spatial domain is recreated.
52
Inverse FFT Original
53
Likewise the intensity of the spots can be
reduced and an inverse FFT will result in the
reduction of the apparent frequencies such as
chatter in a TEM section.
54
Ethics of Digital Image Processing
Although most of what can now be accomplished in
the way of image manipulation was possible by way
of clever darkroom technique it is now much
easier to falsify a digital image.
55
Rules of Image Processing
1. Save the original unaltered image. For your
own sanity you should only process duplicates of
the original and the original image should be
made available for others to examine. 2. Process
only as needed. The removal of artifacts is
acceptable if it does NOT alter the data
contained in the image. Processing for emphasis
(i.e. colorization) is generally acceptable. 3.
Document EACH step. Your materials and methods
should specify precisely which steps were taken
in processing the image.
56
Use of Color in Electron Micrographs
Microheater
Blood Clot
HIV budding from cell
57
Use of Color in Electron Micrographs
Drosophila melanogaster By David Scharf
58
(No Transcript)
59
(No Transcript)