Ezhilarasi.

M et al

Calicut Medical Journal 2008; 6(2):e5 ________________________________________________________________________ Special article

Fractal Geometry in 2D Matrix Array for Realtime 3D Ultrasound Imaging

Ezhilarasi.M*, Rajaram.M**, Sivanandham S.N***
*Assistant Professor, Department of Biomedical Instrumentation Engineering, Avinashilingam University, Coimbatore. **Assistant Professor, Department of Electrical Engineering, Thanthai Periyar Institute of Technology, Vellore. *** Professor and Head, Department of Computer Science Engineering, PSG College of Technology, Coimbatore.

______________________________________________________________________ Abstract
Introduction: Ultrasound (US) imaging has become an effective imaging system in medicine. The US images are generally created in two-dimensional matrix format; however three dimensional images offer better observation and interpretation. Several methods are available to create three-dimensional (3D) images but for direct volume scanning special transducers with two dimensional (2D) matrix arrays should be used. These transducers consist of piezoelectric elements arranged in two dimension matrix form. Methods: Since the construction of such 2D transducers is complex, simulation of these transducers is an appropriate way to analyse and optimize their beam pattern for a particular application. In this research, Field II simulator is used to design and test the transducer. Results: The fractally thinned array proves to be a cheaper and provides good resolution . Conclusions: Many researchers are still working on the proper design of two dimensional array transducers. But a reliable array is yet to be designed. In this research paper, the application of fractal geometry in the design of the ultrasound transducer is demonstrated. Keywords 2D matrix array, Ultrasound, 3D imaging, probe design, fractal array

Introduction
Various methods available for 3D ultrasonography are: free-hand acquisition using a conventional 2D ultrasound transducer without position sensing, freehand acquisition using a conventional 2D ultrasound transducer with position sensing and automated acquisition using dedicated mechanical volume probes. These methods rely on the acquisition of a series of 2D frames using 1D linear arrays and the frames are reassembled by the ultrasound equipment and displayed as a 3D 1,2 volume dataset . In order to produce realtime three-dimensional (4D) ultrasonographic images (i.e., three spatial dimensions plus motion), volume datasets need to be acquired directly and displayed faster than the capacity of the human eye to retain a visual impression. The retaining rate

Page number not for indexing purposes_____________________________________________________1

Ezhilarasi.M et al

Calicut Medical Journal 2008; 6(2):e5 ________________________________________________________________________

of human visual system is estimated as one 3 tenth to one thirtieth of a second . Currently available mechanical probes are capable of volume acquisition rates of up to 40 volumes per second. This is achieved by rapid oscillation of a convex array mounted on a mechanical wobble located inside the transducer housing (rapidly oscillating 4 While this transducer technology) . technology is widely available and currently offers the best spatial resolution, it is not capable of direct volume scanning. Direct volume scanning is a term proposed 3 by Deng, in 2003, to describe 3D/4D systems capable of scanning a volume of interest (i) in its totality, (ii) within a time in which movement is negligible (in an instant), and (iii) with sufficient spatial resolution. But in mechanical transducers the totality of the volume of interest is not scanned in an instant and therefore they are indirect 3 volume scanning methods . This has considerable implications for the examination of the moving fetus and, specially, the fetal heart, since motion artifacts can easily interfere with the quality of the images obtained if the structure of interest moves faster than the speed at which the volume dataset is being acquired. 2D matrix Array In contrast, 2D matrix array transducers allow direct volume scanning by electronic interrogation of a region of interest and acquisition of a pyramidal volume of 59 ultrasonographic data, currently at rates of 24 volumes per second. Although the use of 2D arrays to produce real-time 3D ultrasound images is the most convenient, the current limitation of this acquisition mode is a narrow field of view in the azimuth plane, since only one-third of the matrix arrays aperture is active during acquisition. Need for reduction of elements: One of the primary challenges in realizing a 2D array is that the number of channels (the electronic preprocessing units) required becomes prohibitively large as each element in the array needs separate preprocessing unit. For example, a 6464 element array requires over 4000 channels. For this reason, much research effort has been devoted towards development of methods for reducing the number of elements (called thinning) in a 2D array, while maintaining the beam properties similar to those obtained with the full 2D array.

Methods
Thinning Methods: Only two methods have been adopted so far in thinning of 2D matrix array. One is periodical, where the elements are removed at regular intervals and the other is random, where the elements are removed based on mathematical equations and the elements are removed randomly. Periodic array gives sharper beam at the centre but the energy loss around the centre beam is more. Random array reduces the energy loss around the centre beam and is robust but the main beam is degraded.

Fractal Thinning: What is Fractal? The term was coined by Benot Mandelbrot, the Father of fractal geometry, in 1975 and

Page number not for indexing purposes_____________________________________________________2

Ezhilarasi.M et al

Calicut Medical Journal 2008; 6(2):e5 ________________________________________________________________________

was derived from the Latin word fractus meaning "broken" or "fractured". Flowers, trees, cliffs, clouds and whatever we see in nature, they all have fractal structures. Fractal shapes are not new. Traditional geometry can not describe these structures. How can we describe a cloud-as a circle with fuzzy edges? What can we call a mountain-a triangle with bumps? Close examination of the veins of a leaf reveals a branching pattern similar to the whole tree. Just as the trees overall pattern is carried by its genetic code, so too is a fractals shape carried by its mathematical code, as its seed. When the seed (called the basic pattern) repeats itself (iteration) in a particular fashion, it produces fractal structure.

The Figure 1 shows the seed and its growth results in the formation of fractal structure.

Seed ------------------------------------------------------

Fractal

Fig.1 Fractal generation by mere repletion of a basic pattern

Few other fractal structures generated are shown in Figure 2.

Fig.2 The natural structures created by fractal patterns Fractal array in ultrasound imaging: Fractal array behaves neither like Periodic array and nor like random array but exploits the advantages of both the methods. Fractal array gives out better main beam and robust in nature. Page number not for indexing purposes_____________________________________________________3

Ezhilarasi.M et al

Calicut Medical Journal 2008; 6(2):e5 ________________________________________________________________________

Results:
Need for simulation: since the construction of 2D transducers is complex and expensive, simulation of these transducers is an appropriate way to analyse and optimize their beam pattern. Simulation: Field II is a set of programs for simulating ultrasound transducer fields and ultrasound imaging. The programs are capable of calculating the emitted and pulseecho fields for both the pulsed and continuous wave case for a large number of different transducers. Also any kind of linear imaging can be simulated as well as realistic images of human tissue. The anatomic phantoms are attempts to generate images as they will be seen from real human subjects. This is done by drawing a bitmap image of scattering strength of the region of interest (Figure 3 ;see below). This map then determines the factor multiplied onto the scattering amplitude generated and models the difference in the density and speed of sound perturbations in the tissue. Simulated boundaries were introduced by making lines in the scatterer map along which the strong scatterers were placed. This is marked by completely white lines shown in the scatterer maps. Currently, the elevation direction is merely made by making a 15 mm thickness for the scatter positions, which are randomly distributed in the interval. A phantom for a 3 month old fetus has been made. 200,000 scatterers were randomly distributed within the phantom. The phantom was scanned with a 5 MHz 64 element phased array transducer with lambda/2 spacing and Hanning apodization. A single transmit focus 70 mm from the transducer was used, and focusing during reception is at 40 to 140 mm in 10 mm increments. The images consist of 128 lines with 0.7 degrees between lines and is shown in Figure 4. The images have many of the features from real scan images, but still lack details. This can be ascribed to the low level of details in the bitmap images, and that only a 2D model is used. But the images do show great potential for making powerful fully synthetic phantoms that can be used for image quality evaluation. The phantom was again scanned with a 5 MHz 2D dense array transducer and fractal array transducer with lambda/2 spacing and Hanning apodization. A single transmit focus 70 mm from the transducer was used, and focusing during reception is at 40 to 140 mm in 10 mm increments. The images consist of 128 lines with 0.7 degrees between lines (Figure 5). The phantom was then scanned with the transducer of same specifications except the number of elements which is reduced by 70% by fractal theory (Figure 6)

Page number not for indexing purposes_____________________________________________________4

Ezhilarasi.M et al

Calicut Medical Journal 2008; 6(2):e5 ________________________________________________________________________

Figure 3 2D Phantom bitmap image

0 10 20 Axial distance [mm] 30 40 50 60 70 80 -50 0 Lateral distance [mm] 50

Figure 4 Phantom image scanned with 1D array

Page number not for indexing purposes_____________________________________________________5

Ezhilarasi.M et al

Calicut Medical Journal 2008; 6(2):e5 ________________________________________________________________________

0 10 20 Axial distance [mm] 30 40 50 60 70 80 -50 0 Lateral distance [mm] 50

Figure 5 Phantom image scanned with 2D dense array (no.elements:1804)

0 10 20 Axial distance [mm] 30 40 50 60 70 80 -50 0 Lateral distance [mm] 50

Figure 6 Phantom image scanned with 2D fractal array (no.elements:564 Almost 70% Thinning) Page number not for indexing purposes_____________________________________________________6

Ezhilarasi.M et al

Calicut Medical Journal 2008; 6(2):e5 ________________________________________________________________________

Conclusion:
The images obtained with the 1D array and 2D Dense and fractal arrays prove that the fractal array is capable of reducing the number of channels drastically (the cable is References: 1. Pretorius DH, Borok NN, Coffler MS, Nelson TR. Three-dimensional ultrasound in obstetrics and gynecology. Radiol Clin North Am. 2001;39:499521. 2. Timor-Tritsch IE, Platt LD. Threedimensional ultrasound experience in obstetrics. Curr Opin Obstet Gynecol. 2002;14:56975. 3. Deng J. Terminology of three-dimensional and four-dimensional ultrasound imaging of the fetal heart and other moving body parts. Ultrasound Obstet Gynecol. 2003;22:336 44. 4. Deng J, Rodeck CH. New fetal cardiac imaging techniques. Prenat Diagn. 2004;24:1092103. 5. von Ramm OT, Smith SW. Real time volumetric ultrasound imaging system. J Digit Imaging. 1990;3:26166. 6. von Ramm OT, Smith SW, Pavy HG Jr. High-speed ultrasound volumetric imaging system. II. Parallel processing and image ___________________________________________________________________________ Address for correspondence: M.Ezhilarasi Assistant Professor in Biomedical Instrumentation Engineering, Avinashilingam University for Women, Coimbatore-108 E-Mail:ezhilrasi@hotmail.com display IEEE Transactions on Ultrasonic. Ferroelectrics and Frequency Control. 1991;38:10915. 7. Smith SW, Trahey GE, von Ramm OT. Two-dimensional arrays for medical ultrasound. Ultrason Imaging. 1992;14:213 33. [PubMed] 8. von Ramm OT. 2-D arrays. Ultrasound Med Biol. 2000;26 (Suppl 1):S10S12. [PubMed] 9. Stetten G, Tamburo R. Real-time threedimensional ultrasound methods for shape analysis and visualization. Methods. 2001;25:22130. [Plumed] 10.Ota T, Fleishman CE, Strub M, Stetten G, Ohazama CJ, von Ramm OT, et al. Realtime, three-dimensional echocardiography: feasibility of dynamic right ventricular volume measurement with saline contrast. Am Heart J. 1999;137:95866. [PubMed] the costliest part of Ultrasound machine) while maintaining the image quality. This paves the way for direct collection of volumetric data set and hence the high quality imaging is made possible.

Page number not for indexing purposes_____________________________________________________7