topeeng.gif (8383 bytes)

[ Scientific Activity - Actividad Científica ] [ Brief Communications - Temas Libres ]

Elimination of noise via morphological filters and components labeling. Its use in the study of angiogenesis.

Rodríguez M., Roberto; Castellanos B., Ingrid; Alarcón M., Teresa; (+) Wong N., Roberto ; Felipe R., Edgardo and Sanchez C., Leudis.

Group of Digital Signal Processing (GUPDIS) Institute of Cybernetics, Mathematics and Physics (ICIMAF)
CITMA (+) Department of Phatology, Hospital "Dr. Carlos J. Finlay"
La Habana, CUBA

Abstract
Introduction
Experiments
Conclusions

Abstract
The aim of this work is the elimination of the impulsive noise (IN) resulting of the image segmentation of blood vessels (BV) in the angiogenesis process (AP). We propose an alternative scheme based on components labeling (CL) and morphological filters (MF). Final results were compared with manual segmentation realized by an expert. It is demonstrated by extensive experimentation, using real image data, that proposed strategy is fast and robust in the environment of a personal computer. These images will be subject to a further morphometrical analysis, in order to diagnose and prognosticate automatically malign tumors (MT).
Introduction: In many image processing tasks, segmentation is an important step toward image analysis. It allows quantification and visualization of the objects of interest. Efforts towards the solution of the segmentation problem are motivated by the variety of applications wherein segmentation plays a crucial role.
The main objective of this work is to present the experimental results to eliminate the IN resulting of the BV segmentation in color images of MT in an AP.
Experiments: We decided to use a processing strategy that takes information either from the image resulting of the BV segmentation, and from the image obtained using generalized morphological filters (GMF), as well. Then, this image is filtered with a GMF of large kernel and high strictness whose goal is to eliminate all noise, but any BV can not totally be deleted.
Conclusions: In conclusion, we proposed an alternative scheme using MF and CL to eliminate the IN resulting of the image segmentation of BV in the AP. We can conclude that the use of GMF was better than ordinary ones, although they do not eliminate completely the IN. The experimental results presented in paper demonstrated also the effectiveness of this strategy for filtering of IN.

Top

Introduction: In many image processing tasks, segmentation is an important step toward image analysis. It allows quantification and visualization of the objects of interest. Recently, image segmentation methods were extensively reviewed. It was concluded that segmentation of medical images is so far a difficult task and fully automatic segmentation procedures are far from satisfying to specialists in many realistic situations. Nevertheless, efforts towards the solution of the segmentation problem are motivated by the variety of applications wherein segmentation plays a crucial role.

One of the most common image features used in machine vision are edges, and there is a substantial body of research on various techniques for performing edge detection. A drawback with using edges is that not only do edge detectors extract meaningful and useful edges, but also many other spurious ones which arise from noise and minor changes in intensity values. When the intensity or structure of the objects differs significantly from the surroundings, segmentation is obvious. In addition, in all other situations manual tracing of the object boundaries by an expert seems to be only valid truth, but it is undoubtedly a very time-consuming task.

Many methods have been proposed so far for the elimination of impulsive noise, among them, the median filter , rank-order filter , and more recently the so-called morphological filters.

Mathematical morphology (MM) is a powerful tool inside the DIP, since that science can rigorously quantify many aspects of the geometrical structure of the images in a way that agrees with the human intuition and perception . This method developed in the 1960s mainly by Serra and Matheron has its mathematical origins  in concepts of the set theory, convex analysis, integral geometry, stereology and geometrical probabilities.

The main objective of this work is to present the experimental results to eliminate the impulsive noise resulting of the blood vessels segmentation in color images of malign tumors in an angiogenesis process, using a combination of morphological filters and components labeling. Performance comparison of this method is made by computing the percent of root-mean-squared difference between the filtering results and the ideal segmentation by hand realized by an expert. Difference errors less than 3% were observed. It is demonstrated by extensive experimentation, using real image data, that proposed strategy here is fast and robust in the environment of a personal computer.

Top

Experiment: Figure 1 shows the segmentation scheme proposed by our. Figure 2a shows an example of a image representing a histological cut of malign tumors of soft parts, where the blood vessels (BV) are shown. The segmentation results according to the processing strategy presented can clearly be observed in figure 2b. It can be noted that, although the BV are well defined, the image presents a considerable noise.

fig1.gif (1810 bytes)

 

fig2.gif (9180 bytes)

 

Top

Figure 3 shows a block-schema of the method used to eliminate the impulsive noise. It is important to point out that the most significant step in this strategy is the filtering process to eliminate completely the noise of the segmented image, although the BV are affected in some measure, all of them should appear in the filtered image; otherwise in the classification process of the labels, some BV would be eliminated. According to the strategy proposed, the impulsive noise in all images always was eliminated, since it was of lesser size than the BV.

fig3.gif (2027 bytes)

 

The proposed strategy offered us very good results, since images completely free of impulsive noise were obtained, keeping in the BV its size and shape. Figure 4 shows the effect of this operation. Note in the figure 4c the similarity of BV compared with the manually segmented image shown in figure 4d.

fig4.gif (7698 bytes)

Conclusions: In conclusion, we proposed an alternative scheme using morphological filters and components labeling to eliminate the impulsive noise resulting of the image segmentation of blood vessels in the angiogenesis process. We can conclude that the use of generalized morphological filters was better than ordinary ones, although they do not eliminate completely the impulsive noise. The experimental results presented above demonstrated also the effectiveness of this strategy for filtering of impulsive noise. We demonstrated by extensive experimentation, using real image data, that proposed scheme is fast and robust in the environment of a personal computer. These images will be subject to a further morphometrical analysis, in order to diagnose and prognosticate automatically malign tumors.

Questions, contributions and commentaries to the Authors: send an e-mail message (up to 15 lines, without attachments) to image-pcvc@pcvc.sminter.com.ar , written either in English, Spanish, or Portuguese.

Top


© CETIFAC
Bioengineering
UNER

Update
Dic/03/1999