75
Edge Detection and Segmentation of Images
A typical solution for these problems is dividing the original image into some
subimages in such a way that the histogram of each subimage can be easily separated
into two parts with a simple thresholding process. This means that for each subimage one must select a suitable threshold to segment the histogram of that subimage
into two parts. Then, the segmented subimages are put together to form the overall
segmented image. The bottleneck of this method is designing a reasonable process to
divide the original image into subimages. Selecting threshold values for the resulting
subimages is another issue to be dealt with.
4.3.3.2 Region Growing
The methods introduced so far belonged to the first category of segmentation algorithms. These methods are based on finding differences and boundaries between
parts of an image. In other words, the methods discussed earlier use discontinuities
among gray levels of entities in an image (e.g., point, line, edge, and region) to segment different parts of the image. In this section, we focus on the second category of
segmentation techniques that attempt to find segmented regions of the image using
the similarities of the points inside image regions.
In region growing methods, segmentation often starts by selecting a seed pixel
for each region in the image. Seed pixels are often chosen close to the center of the
region or object. For example, if we are to segment a tumor from the background, it
is always advisable to select the seed point for the tumor in the middle of the tumor
and the seed point for the background somewhere deep in the background region.
Then, the region growing algorithm expands each region based on a criterion, which
is defined to determine similarity between pixels of each region. This means that
starting from the seed points and using the criterion, algorithm decides whether the
neighboring pixels are similar enough to the other points in the region, and if so,
these neighboring pixels are assigned to the same region that the seed point belongs
to. This process is performed on every pixel in each region until all the points in the
image are covered.
The most important factors in region growing are selecting a suitable similarity
criterion and starting from a suitable set of seed points. Selecting similarity criteria
mainly depends on the type of the application in hand. For example, for the monochrome (gray level) images, similarity criterion is often based on the gray-level features and spatial properties such as moments or textures.
Example 4.7
In this example, a subimage shown in Figure 4.11a is segmented through growing
of the seed points for the two regions shown in Figure 4.11a. The range of gray
level in this image is from 0 to 8. As can be seen, the gray level of each pixel is
also shown in Figure 4.3a, and the seed points are marked by red-underlined gray
levels. In this example, the similarity criterion is defined as follows: two neighboring pixels (i.e., horizontal, vertically, or diagonally neighboring pixels) belong to
the same region if the absolute difference of their gray levels is less than T = 3.
With this criterion and starting the identified seeds in Figure 4.11a, the segmented
image of Figure 4.11b is obtained.
Edge Detection and Segmentation of Images
A typical solution for these problems is dividing the original image into some
subimages in such a way that the histogram of each subimage can be easily separated
into two parts with a simple thresholding process. This means that for each subimage one must select a suitable threshold to segment the histogram of that subimage
into two parts. Then, the segmented subimages are put together to form the overall
segmented image. The bottleneck of this method is designing a reasonable process to
divide the original image into subimages. Selecting threshold values for the resulting
subimages is another issue to be dealt with.
4.3.3.2 Region Growing
The methods introduced so far belonged to the first category of segmentation algorithms. These methods are based on finding differences and boundaries between
parts of an image. In other words, the methods discussed earlier use discontinuities
among gray levels of entities in an image (e.g., point, line, edge, and region) to segment different parts of the image. In this section, we focus on the second category of
segmentation techniques that attempt to find segmented regions of the image using
the similarities of the points inside image regions.
In region growing methods, segmentation often starts by selecting a seed pixel
for each region in the image. Seed pixels are often chosen close to the center of the
region or object. For example, if we are to segment a tumor from the background, it
is always advisable to select the seed point for the tumor in the middle of the tumor
and the seed point for the background somewhere deep in the background region.
Then, the region growing algorithm expands each region based on a criterion, which
is defined to determine similarity between pixels of each region. This means that
starting from the seed points and using the criterion, algorithm decides whether the
neighboring pixels are similar enough to the other points in the region, and if so,
these neighboring pixels are assigned to the same region that the seed point belongs
to. This process is performed on every pixel in each region until all the points in the
image are covered.
The most important factors in region growing are selecting a suitable similarity
criterion and starting from a suitable set of seed points. Selecting similarity criteria
mainly depends on the type of the application in hand. For example, for the monochrome (gray level) images, similarity criterion is often based on the gray-level features and spatial properties such as moments or textures.
Example 4.7
In this example, a subimage shown in Figure 4.11a is segmented through growing
of the seed points for the two regions shown in Figure 4.11a. The range of gray
level in this image is from 0 to 8. As can be seen, the gray level of each pixel is
also shown in Figure 4.3a, and the seed points are marked by red-underlined gray
levels. In this example, the similarity criterion is defined as follows: two neighboring pixels (i.e., horizontal, vertically, or diagonally neighboring pixels) belong to
the same region if the absolute difference of their gray levels is less than T = 3.
With this criterion and starting the identified seeds in Figure 4.11a, the segmented
image of Figure 4.11b is obtained.