41
Image Filtering, Enhancement, and Restoration
s = T(r)
s
s 2
s 1
r
r 1
r 2
FIGURE 3.3 Contrast enhancement using point processing.
3.2.1 CONTRAST ENHANCEMENT
If the object of interest (to be analyzed) occupies only a specific range of the gray
scale, then one may want to manipulate the image such that the object occupies a
larger range of the gray level and therefore increase the visibility of the object. In
medical images, for example, often one needs to analyze a tumor surrounded by
organs that are darker or brighter than the tumor is. If other organs and objects are
not the main focus of the analysis, then in order to visualize the tumor better, one can
perform contrast enhancement to “stretch” the gray level of the tumor.
Contrast enhancement is a method to create better visibility of a particular range
of gray level that corresponds to the object to be studied. This stretch of gray-level
range is illustrated in Figure 3.3.
As can be seen in Figure 3.3, the exact shape of the transformation applied for
contrast stretching is controlled by values of (r 1 , s 1 ) and (r 2 , s 2 ). It is apparent that the
choice of r 1 = s 1 and r 2 = s 2 would reduce the transformation to a linear operation that
has no effect on the gray level of the original image. In practice, we often select these
values such that the interval [s 1 , s 2 ] covers the gray-level range of the object of interest
and the interval [r 1 , r 2 ] provides the desired range of gray level for a better visibility of
the object in the target image. This implies that for a better object visibility “r 2 − r 1 ”
must be much larger than “s 2 − s 1 ”; in other words, with this condition, the differences among the gray levels in the region of interest are amplified and enhanced in
the target image.
Next, we show how to implement point processing in MATLAB ® .
Example 3.1
In this example, we enhance and stretch the gray level of the original image in
the interval of [150, 200] to the desired interval of [105, 200]. MATLAB code of
this example is shown in the following text. First, we read the original image with
“imread” command. This image shows that ultrasonic image of the heart and its
compartments. We have intentionally chosen a noisy ultrasonic image to better
represent the quality of images to be processed in many medical applications.
Image Filtering, Enhancement, and Restoration
s = T(r)
s
s 2
s 1
r
r 1
r 2
FIGURE 3.3 Contrast enhancement using point processing.
3.2.1 CONTRAST ENHANCEMENT
If the object of interest (to be analyzed) occupies only a specific range of the gray
scale, then one may want to manipulate the image such that the object occupies a
larger range of the gray level and therefore increase the visibility of the object. In
medical images, for example, often one needs to analyze a tumor surrounded by
organs that are darker or brighter than the tumor is. If other organs and objects are
not the main focus of the analysis, then in order to visualize the tumor better, one can
perform contrast enhancement to “stretch” the gray level of the tumor.
Contrast enhancement is a method to create better visibility of a particular range
of gray level that corresponds to the object to be studied. This stretch of gray-level
range is illustrated in Figure 3.3.
As can be seen in Figure 3.3, the exact shape of the transformation applied for
contrast stretching is controlled by values of (r 1 , s 1 ) and (r 2 , s 2 ). It is apparent that the
choice of r 1 = s 1 and r 2 = s 2 would reduce the transformation to a linear operation that
has no effect on the gray level of the original image. In practice, we often select these
values such that the interval [s 1 , s 2 ] covers the gray-level range of the object of interest
and the interval [r 1 , r 2 ] provides the desired range of gray level for a better visibility of
the object in the target image. This implies that for a better object visibility “r 2 − r 1 ”
must be much larger than “s 2 − s 1 ”; in other words, with this condition, the differences among the gray levels in the region of interest are amplified and enhanced in
the target image.
Next, we show how to implement point processing in MATLAB ® .
Example 3.1
In this example, we enhance and stretch the gray level of the original image in
the interval of [150, 200] to the desired interval of [105, 200]. MATLAB code of
this example is shown in the following text. First, we read the original image with
“imread” command. This image shows that ultrasonic image of the heart and its
compartments. We have intentionally chosen a noisy ultrasonic image to better
represent the quality of images to be processed in many medical applications.