0% found this document useful (0 votes)
31 views2 pages

Software Development For Low Cost, High Quality, Real-Time 4D Ultrasound Imaging System On Personal Computers (Pes)

Download as pdf or txt
Download as pdf or txt
Download as pdf or txt
You are on page 1/ 2

~ 26

t h
NATIONAL RADIO SCIENCE CONFERENCE (NRSC2009)
~
March 17-19,2009, Faculty of Engineering, Future Univ., Egypt
I P13 ~
Software Development for Low Cost, High Quality, Real-time 4D
Ultrasound Imaging System on Personal Computers (pes)
Marwan M. Abd Ellah
1. Introduction
Alaa' A. Megawer
marwanabdellah@yahoo. com
alaamegawer2@yahoo. com
ymk@k-space. org
Dr. Yasser M. Kadah
Ultrasound imaging is considered one of the most important imaging modalities used widely in the
field of biomedical imaging; it has broad latitude of clinical and diagnostic applications. Real-time 4D
ultrasound imaging is a medical technique used in obstetrical scans, which provide sonographers with rapidly
animated three-dimensional images of the fetus. Sonographers around the world were always dreaming with
these animated three dimensional images instead of ordinary B-mode two-dimensional scans. They were
deliberating all the times in interactive fetus visualization. This interactive visualization will lead to better
diagnosis and consequently to the welfare of the community.
2. Problem definition
Real-time ultrasound imaging modality is the system which can meet sonographers' dreams. However,
until recently it was not possible to develop this type of ultrasound machines at all. This was down to the lack of
adequate hardware and software platforms that make the developers capable of implementing interactive volume
reconstruction. Despite hardware availability in that time, this hardware was not widespread in the market, as
well as components' cost was too expensive. Consequently, ultrasound systems' overall cost was very expensive.
Besides that, software development was a tedious matter in the past. This was a consequence of the limited
functionality of the fixed-function pipeline on the available graphics cards in that time. Explicitly speaking, the
real problem in an interactive 3D ultrasound lies in the massive data sets to be processed and rendered per
second. All of these reasons made the developers waiting for affordable and flexible hardware that can make
them capable of developing and manufacturing low cost and high performance 4D ultrasound machines that
support interactive 3D volume visualization.
3. Technology
After the advent of the current state-of-the-art multi-core GPUs (Graphics processing units) in this era,
graphics cards capabilities became more and more and even more flexible to be utilized in interactive
visualization, they offer a level of programmability and performance that not only makes it possible to perform
traditional workstation tasks on a cheap personal computer, but even enables the use of highly complicated
rendering algorithms that previously could not be employed for real-time visualization at all. These modem
GPUs are very brilliant and efficient at manipulating and displaying massive computer graphics datasets, as well
as their highly parallel structure makes them more effective than general-purpose CPUs for a range of complex
algorithms. These GPUs' pipeline is now programmable - yes they are programmable - not fixed-function any
more. This programmability capability is the real power that has lead to great improvements in the field of
medical interactive visualization. Thanks to great evolution of the consumer graphics hardware, we can exploit
the nowadays available low cost and high performance hardware great functionality to write the adequate
software that can yield the sonographers' dream; it is high quality interactive visualization that combines real-
time performance with high resolution imaging.
4. Our work
Our target in this project is to develop applicable software dedicated for low cost, high quality, and
high performance real-time 4D ultrasound machines. This software must have the flexibility to run on the
current affordable consumers' graphics hardware available in the global market. As well as it must achieve what
we are originally seeking for; the interactive visualization or real-time performance.
26
th
NATIONAL RADIOSCIENCE CONFERENCE, NRSC'2009
Future University, 5
th
Compound, NewCairo, Egypt, March17- 19,2009
March 17-19,2009, Faculty of Engineering, Future Univ., Egypt
26
t h
NATIONAL RADIO SCIENCE CONFERENCE (NRSC2009)
~
I P13 Q]
Regarding our choice for a hardware platform, we set on the nvidia GeForce 9800 GS graphics card
which is considered a suitable graphics platform we can trust to implement our system on. This graphics card
has the fol1owing features: 64 Stream Processors, 530 MHz core clock, 256 bit memory interface, 51.2 GB/s
memory bandwidth, 17.0 billion texels/s texture fill rate. This GPU pipeline is programmable with both of its
vertex and fragment processors. Shaders are codes written to program the unit (GPU) programmable rendering
pipeline. We are going to use the Cg (C for graphics) shading language to write our vertex and fragment
shaders. OpenGL will be used as the 3D graphics API (application programming interface) or graphics pipeline
model and the interface between CPU and GPU. Both of the Cg and OpenGL codes will be encapsulated in the
MS Visual C++ programming languages classes. And finally GUl design will be implemented using the QT
library. The next figure depicts the pipeline we are going to follow.
~ ~ ~ l
Core2 Duo :
Processor I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I _________________________ J
CPU-GPU Boundary (PCI-e)
GPU Front-e nd
Tex ture Memory
Display
VGAOut
Frame Buff er
Nvidia GeForce
9800 Programmable
Pipeline
~ _J - - - - - - - - - - - - - - - -- - - - - - - - - - - -- -
Among the versatile volume rendering techniques, the need to a real optimized one that exploit the
flexible programming model and 3D texturing affordable capabilities of modem GPUs is crucial such that it will
final1y affect the interactivity of the visualized volume. Although it is possible to implement other popular
volume rendering algorithms on the GPU, such as ray casting or others, we are going to use the most one that
offer a great level of interactivity without sacrificing the quality of rendering. It is texture-based volume
rendering.
5. Conclusion
Interactive volume visualization was a dream that can't easily and cheaply be achieved in the past. This
was due to the lack of the high performance graphics hardware that capable us of developing real-time 4D
ultrasound machines. Current state-of-the-art low cost consumer graphics chips afford a level of
programmability and performance that makes it possible to easily develop interactive volume visualization on
them. Processing powers of the these chips has increased so much that today we can get real-time ultrasound
equipment to acquire and display the 3D datasets with their multi-planar reformations and renderings in real
time. In this project we are going to exploit the great programming capability of one of the available nvidia
graphics cards chips to implement a texture-based volume rendering techniques on it, seeking for developing
applicable software that achieve real-time and interactive ultrasound volume visualization.
Writing software is an easy job, but exploiting the ultimate hardware capabilities for developing bug-
less (as possible) and suitable software for real-time performance is the real hard target we are trying to achieve
in this project. Reaching this target in this project will give us the spark to put our efforts towards a final
product ; it is Real-time 4D ultrasound machine.
26
th
NAnONAL RADIO SCIENCECONFERENCE, NRSC'2009
FutureUniversity, 5
th
Compound, NewCairo, Egypt, March 17- 19,2009

You might also like