Academic Open Internet Journal

www.acadjournal.com

Volume 14, 2005

 

 

 

THE ROLE OF ACTIVE CONTOUR MODELS

IN

BIOMEDICAL IMAGE  ANALYSIS

 

H.S.Sheshadri (research scholar) & Dr A.Kandaswamy, prof & Dean,

 Dept  of  ECE , PSG College of Technology, Coimbatore.641004.

 

 

Abstract.

Medical imaging allows scientists and physicians to decide about life saving information regard to the human physiological activities. The role of medical imaging in diagnostic features is innumerable and the computer assisted imaging is now the challenging work of researchers. This paper gives an over all view about the use of active contour models in the diagnosis and analysis of biomedical images such as X-ray, MRI, Mammograms etc..

      Deformable models are the active contour models used for the identification of the geometric size and shape of abnormal growth of any organs, tumors, lesions etc. as   identified by any of the medical imaging modalities. It deals with the physics, geometry and approximation theory and finally gives the exact place, position and size of openings required to perform any surgery. Generally SNAKES are the deformable models that are widely being used in medical image analysis and hence in this paper an attempt has been made to focus on the application of snakes in biomedical image analysis. The active contour model software has been employed on the images such as tumors, lesions etc. and the results have shown better performance in the detection of abnormalities. A brief details of the various models and their applications with respect to biomedical image analysis are also  discussed.

 

Keywords.  Deformable models,. Active contour models, Segmentation, Shape modeling.

 

MATHEMATICAL  MODEL OF A SNAKE:

 

General model of an active contour model in the image plane(x,y,) is given by

 where x and y are coordinate functions and  is the parametric domain.The shape of the contour subject to an image I(x,y,) is dictated by the functional

.

 

The functional can be seen as energy representation and the final shape of the contour corresponds to the minimum of this energy.The mathematical equation may also be written as

 

It characterizes the deformation of a flexible contour. The above model has two parameters. One w1(s) controls the tension and w2(s) controls the rigidity. In general

.

Where p(x,y) denotes a scalar potential function defined on image plane. To apply snakes to images ,external potentials are designed whose local minima coincide with the intensity extreme edges and other image feature of interest.

In accordance to the theory of calculations of variations we have Euler _Lagrange equation as

.

This vector differential equation  gives the condition for balance of vector  forces under equilibrium. The first two terms represent the internal stretching and bending foresees where as the third term represent the external force that couple the snake to the image data.

 

Medical image  analysis with contour models.

 

    Active contour models are being employed to images generated by imaging mdalities  such such as Xray,CT, an Angiogram,MRI, and ultrasound .Two and three dimensional contour models are being used. Here we discuss the use of snakes as applied to ultrasound and X-ray images (two dimensional only).They are being used to  segment , visualize ,track and quantify any anatomical  structures. Snakes are being used to identify the abnormalities in the structures of brain, heart, blood vessels, kidney, lungs stomach, and liver skull and also in cellular  structures such as neurons and chromosomes. In the following section we discuss some typical applications of snakes for segmentation and matching.

  Most of the image segmentation tools employ the snakes for the contour analysis which employ manual slicec editing. But manual slice editing suffers from several drawbacks such as operator bias, false edge detection, artifacts  due to clicking the mouse etc...These drawbacks can  be solved by employing automatic  or user friendly software on snakes.

 

Methodology:

 

 In this paper we present a detailed method to use an active contour model to visulaise 2d images. The images that are taken as pilot study are of CT images, X_ray images  and MRI images. The  results obtained after segmentation and contour detection are as shown in figs. The following are the steps employed for using the snake model software.

1.      Read the image

2.      Perform   preprocessing..(filtering/enhancement/)

3.      Apply segmentation process.

4.      Identify the region of interest.

5.      Apply  contour model software

6.      Get the detailed information about  the geometry of the image.

7.      Repeat the above procedure in case of multiple contour analysis.(once for each of region of interest).

8.      Selection of various parameters depends on the accuracy required.

 

We employ here simple technique of segmentation using available commands in image processing toolbox of Matlab (v6.1).The SNAKE software has been  used to determine the contour.

 

 

Results: Here we discuss certain applications on the use of active contour models for the analysis of  Ultrasound and MRI images..The figs show the images with the contour model applied, and the various  paramerers are also given for ease of reference.

 

Fig(1),  Contour of a Fibroenchyma                                   Fif(2),contour of a Breast Cyst.

                                                          

 

Fig(3) Chest image                                                            Fig(4)MRI  of heart.

     

 

 

       

                                                                                     

Conclusions: The active contour models play a vital role in determining the exact shape and size of any abnormalities found in any of the biomedical imaging modalities. Here only a qualitative analysis has been given and this topic is still open for a lot of research on the biomedical applications.The  models can be further developed for the analysis of 3d images. Also  research work is  in progress on the use of snakes on color and moving pictures.

      A large number of software packages are now available to perform contour analysis, but  one of the most widely used  is  the SNAKE model as supported by

 Dejan Tomazevic.

 BIPROG, Faculty of Electrical Engineering, University of Ljubljana.

 

 

References.

  1. C. Xu and J. L. Prince, "Snakes, Shapes, and Gradient Vector Flow," IEEE Transactions on Image Processing, 7(3), pp. 359-369, March 1998 (JHU-ECE TR96-15).
  2. D. L. Pham, J. L. Prince, A. P. Dagher, and C. Xu, "An Automated Technique for Statistical Characterization of Brain Tissues in Magnetic Resonance Imaging," International Journal of Pattern Recognition and Artificial Intelligence, 11(8), pp. 1189-1211, 1997
  3. C. Xu and J. L. Prince, "Generalized Gradient Vector Flow External Forces for Active Contours," Signal Processing --- An International Journal, 71(2), pp. 131-139, December 1998.
  4.  Xu, D. L. Pham, M. E. Rettmann, D. N. Yu, and J. L. Prince, "Reconstruction of the Human Cerebral Cortex from Magnetic Resonance Images," IEEE Transactions on Medical Imaging, 18(6), pp. 467-480, June, 1999.
  5. M. E. Rettmann, X. Han, C. Xu, and J. L. Prince, "Automated Sulcal Segmentation Using Watersheds on the Cortical Surface," NeuroImage, Vol. 15, No. 2, pp. 329-344, Feb. 2002.
  6. X. Han, C. Xu, U. Braga-Neto, and J. L. Prince, "Topology Correction in Brain Cortex Segmentation Using a Multiscale, Graph-based Algorithm," IEEE Transactions on Medical Imaging, Vol. 21, No. 2, pp. 109-121, Feb. 2002. 

 

Technical College - Bourgas,

All rights reserved, © March, 2000