73
Edge Detection and Segmentation of Images
(a)
(c)
(b)
(d)
FIGURE 4.9 (a) Original image, (b) image after horizontal line detection, (c) image after
vertical line detection, and (d) image after 45° line detection. (Courtesy of Andre D’Avila,
MD, Heart Institute (InCor), University of Sao Paulo, Medical School, Sao Paulo, Brazil.)
processing because in a typical medical image analysis, one needs to detect regions
representing objects, such as tumors, from the background. Next, we describe two
general methods for region and object detection.
4.3.3.1 Region Segmentation Using Luminance Thresholding
In many biomedical images, the pixels in the objects of interest have gray levels that
are either greater or smaller than the gray levels of the background pixels. In these
images, one can simply extract the objects of interest from the background using the
differences in the gray level. When the object is bright and the background is dark
(or vice versa), separating the interested object from the image can be performed
with a simple thresholding of histogram as described in the following.
Figure 4.10 shows a synthetic cell image that contains cells that are much
darker than the bright background. Also, Figure 4.10 shows the histogram of this,
which shows two almost entirely separate peaks and intervals in the histogram.
Edge Detection and Segmentation of Images
(a)
(c)
(b)
(d)
FIGURE 4.9 (a) Original image, (b) image after horizontal line detection, (c) image after
vertical line detection, and (d) image after 45° line detection. (Courtesy of Andre D’Avila,
MD, Heart Institute (InCor), University of Sao Paulo, Medical School, Sao Paulo, Brazil.)
processing because in a typical medical image analysis, one needs to detect regions
representing objects, such as tumors, from the background. Next, we describe two
general methods for region and object detection.
4.3.3.1 Region Segmentation Using Luminance Thresholding
In many biomedical images, the pixels in the objects of interest have gray levels that
are either greater or smaller than the gray levels of the background pixels. In these
images, one can simply extract the objects of interest from the background using the
differences in the gray level. When the object is bright and the background is dark
(or vice versa), separating the interested object from the image can be performed
with a simple thresholding of histogram as described in the following.
Figure 4.10 shows a synthetic cell image that contains cells that are much
darker than the bright background. Also, Figure 4.10 shows the histogram of this,
which shows two almost entirely separate peaks and intervals in the histogram.