Image Manipulation and Analysis Techniques
Image manipulation and analysis techniques can be classified as follows:
1. Image enhancement and filtering.
This includes:
|
Filtering techniques are divided into two categories:
convolution filters (linear filters) and non-convolution (nonlinear) filters.
Both techniques accomplish their results by examining and processing an
image in small regions, called pixel "neighborhoods." A neighborhood is
a square region of image pixels, typically 3x3, 5x5 or 7x7 in size.
Convolution filters
Example: Hi-Pass. The Hi-Pass filter
accentuates intensity changes in an image by modifying a pixel's value
to exaggerate its intensity difference from its neighbors. It produces
an image with harsh intensity transitions, and generally results in an
image with only edges of high contrast visible. Fine detail with low contrast
is usually lost to the background. This filter can be used when you need
to pull out just the elements having high contrast to the image background.
 Original image |
 Hi-Pass applied |
Non-convolution filters
Example: Erode and Dilate. The Erosion
filter is a morphological filter that changes the shape of objects in an
image by eroding (reducing) the boundaries of bright objects, and enlarging
the boundaries of dark ones. It is often used to reduce, or eliminate,
small bright objects. The Dilation filter is a morphological filter that
changes the shape of objects in an image by dilating (enlarging) the boundaries
of bright objects, and reducing the boundaries of dark ones. The dilation
filter can be used to increase the size of small bright objects.
|
Original Image |
 Erosion Applied |
Image adding and subtracting
Image addition
One image is added to another one, pixel by pixel.
If they are two different images, a blend of the two will result.
If they are identical, the resulting image will have twice the brightness.
This increased brighness is useful to eliminate noise.
Image subtraction
One image can be subtracted from another one, pixel by pixel. Where the pixels have the same value, the resulting pixel is 0. Where the two pixels are different, the resulting pixel is the difference between them.
Example: Image Subtracting
 Image 1 |
 Image 2 |
 Image 3 |
Merging images
Image merging can be done in several ways: it
is possible to extract one colour channel ( e.g. red) and then combine
this with another image. Often the problem arises that the background of
the image that is phased in another one, totally obscures the features
of the recipient image. If the background is flat (i.e. has constant grey
values), the function quad (quadtree) solves this problem. The quadtree
function splits the image up in regions, and examines whether any region
is a uniform. If it is, it is not extracted, if it is not uniform, the
non-uniform part is extracted. This extraction procedure omits the background
so that there is no problem with the background of the extracted image
obscuring what is in the recipient image.
 Image 1 |
 Image 2 |
 Combine image, Using Quadtree |
Frequency domain filtering involves frequency
domain transforms. These transforms change an image from its spatial-domain
form of brightnesses to a frequency domain of fundamental frequency components.
One of the most commonly used is the Fast Fourier Transform. When an image
is transformed with Fast Fourier, and the Fourier frequency is displayed,
it appears symmetrical about the centre. The centre is the zero frequency
point. Two axes run through the centre: the horizontal axis defines the
horizontal (x value) frequency, the vertical axis the vertical (y value)
frequency. The frequency magnitude is determined by the brightness of the
pixel at a particular point. Fast Fourier Transforms are very good for
filtering periodic noise in an image: one can eliminate bright spots in
the Fast Fourier frequency display representing the noise so that the image
is cleared up.
|
Original Image |
 Display of Fast Fourier Transform frequencies |
Image edge enhancement reduces an image to show
only its edge details. It is similar to Hi-Pass, although it focusses more
on the edge itself rather than on the contrast between object and its surroundings.
The most common is Laplacian edge enhancement. It highlights the edges
in an image, irrespective of their orientation.
|
Original Image |
 Laplacian Edge enhancement applied |
Opening and Closing: The opening filter
performs an erosion, then a dilation (see above). In images containing
bright objects on a dark background, the opening filter smoothes object
contours, breaks (opens) narrow connections, eliminates minor protrusions
and removes small dark spots. In images with dark objects on a bright background,
the opening filter fills narrow gaps between objects. The closing
filter is a morphological filter that performs a dilation followed
by an erosion. In images containing dark objects on a bright background,
the opening filter smoothes object contours, breaks narrow connections,
eliminates minor protrusions and removes small bright spots. In images
with bright objects on a dark background, the closing filter fills narrow
gaps between objects.
 Image with background impurities (small black dots) |
 After applying "Closing" filter: impurities have been removed and some connections (eg in images with holes) are about to be broken |
Image Measurements and Feature Extractions
The most common measurements are those that
count objects and/or describe the shape of objects in an image. In Image
Pro it can be found under Measure>Count/Size (except Length which is found
under Measure>Measurements). Image analysis programs offer many different
measurements. The most common are: Area: the pixel area
of the interior of the object.
Perimeter: the pixel
distance around the circumference of the object.
Area to perimeter ratio: a
measure of the object's roundness, or compactness, giving a value between
0 and 1
Major axis: the x,
y endpoints of the longest line that can be drawn through the object
Minor axis: the x,
y endpoints of the longest line that can be drawn through the object while
maintaining perpendicularity with the major axis.
Number of holes: a
count of how many holes exist within the interior of an object.
It has to be noted that these measurements
would normally be expressed in pixels. However, they can be converted to
another unit, such as microns, millimetres, or miles. (In Image pro one
can go to Image>Calibration>Spatial and set number of pixels to another
unit, e.g. microns, or miles). In NIH-Image go to Analyse>Set Scale to
set the number of pixels to another unit.
Many of these measurements add up to so-called feature
segmentation techniques, enabling discrimination between objects of interest.
For example: Image Pro has a function called auto-classification
that uses classifiers such as area, perimeter, major and minor axes to
discriminate between objects. It uses a maximum of three classifiers from
a range of possible ones and can be found in Measure>Count/Size>Measure>Auto-Classification.
Object 1 shape Measures (in pixels)
Perimeter: 871
Area: 30,760
Major axis angle:13 degrees
Major axis width: 360
Minor Axis Width: 152
Object 2 shape Measures (in pixels)
Perimeter: 430
Area: 6455
Major axis angel 0 degrees
Major axis width: 136
Minor Axis Width: 74
|
|
|
These numbers can then be analysed in several ways: by producing a diversity index of each contour and comparing these indices, or by extracting statistical functions called Moment Invariant Functions from the x y coordinates that make up the contour. These include functions such as means, variance, skewness and curtosis. A third method is to describe the contour by Elliptical Fourier Transforms, where a series of shifting ellipses are fitted to the contour with a complex shape (approximating the contour) being generated by the interactive combination of all ellipses used. The shape of these ellipses themselves can then be expresses by a set of numbers, which then, in effect describe the contour of the object itself.
For more information, contact:
Email iaaf@iaaf.uwa.edu.au
Tel: 6488 8649 Fax: 6488 1051
