|
|
|
|
_____________________________________________________________________________
|
Medicine is doubtless one of the branches of science most favoured by digital image processing (DIP). Within medicine, perhaps, the pathological anatomy is a speciality, where the use of different techniques of DIP allows improving the accuracy of diagnosis.
Angiogenesis process is an important aspect in the histopathological research of diseases. This process is studied by pathological anatomy and it is characterised by new formation of blood vessels in the tissue, which increases its growing. Angiogenesis is a normal process that occurs in all tissues, but it is known that in pathological cases the increase of blood vessel is abnormal. For example in mammary, prostate, bladder and brain cancers the greater the number of blood vessels, the more developed is the disease.
In the study of the angiogenesis process the pathologists analyse all information related to the blood vessels using microscopy. This work is very tedious and time consuming. Obviously, automating the analysis is highly desirable. In such a sense, a useful task for DIP should be the segmentation of blood vessel.
The analysis of an image begins with a segmentation process, which differentiates meaningful regions of interest from the background. This step is critical because its outcome serves as the basis for all subsequent analyses, such as extraction of shape features, and finally the interpretation of the scene.
Several segmentation schemes have been proposed to extract objects from clinical images. Among these are the popular thresholding techniques, the fuzzy c means techniques, algorithms using texture and neural network, watershed- based segmentation, and cluster analysis among others. However, a robust segmentation method for all applications is still lacking. So, segmentation process is very dependent on application.
In our case we need to segment colour images for isolating blood vessels. Segmentation of blood vessels is a very difficult task due both to complexity of shape of these structures and to problems inherent to microscopy.
These images will be subject to a further morphometrical analysis, in order to diagnose and prognosticate automatically malign tumors.
|
|
|
Angiogenesis process. It can be observed the blood vessels.
|
|
