Medical Imaging
- Overview
- Challenges in Medical Imaging
- FPOAs in Medical Imaging
- Libraries & IP Cores
- Application Assessments
Attend Free On-Demand Webinar
Implementing the High Profiles of MPEG-2 & MPEG-4/ H.264 on an FPOA Architecture
Medical Image Processing
The Arrix Family of FPOAs are designed to excel at today's most demanding high definition image processing applications. MathStar's 1 GHz FPOA is particularly suited for medical image processing.
- MPEG-2 Encoder
- MPEG-2 Encoder Transport Stream

- JPEG2000 Encoder
- Sobel Edge Detection
- CT Filtered Back Projection
- Digital Beamforming
- 2D Convolution Filter
- Smith-Waterman
- Multi-tap FIR Filter
- H.264 Encoder
- 2D Image Processing
Medical Imaging Overview
Medical imaging modalities such as computed tomography and diagnostic ultrasound are experiencing a period of rapid innovation and use across an expanding breadth of clinical applications, driven by the emergence of 3D / 4D imaging and sophisticated post-processing algorithms. These advancements are driving exponential increases in computational requirements.
Challenges in Medical Imaging
By 2009, 256-slice CT machines will be widely deployed, with 512-slice machines on the horizon, offering the promise of clinically viable analysis of a beating heart. In diagnostic ultrasound, archived volumetric data sets enable clinicians to overcome the inherent limitations of handheld sonography by enabling post processed searches for desired 2D views. Image analysis—for example, image registration to measure the before and after effects of a tumor drug therapy--presents some of the most computationally intensive, and algorithmically complex, challenges in medical diagnostics. Finally, the challenges of archiving data sets are that are growing exponentially in size are now being addressed computationally intense compression techniques such JPEG2000 and MPEG4/H.264.
How Can FPOAs Solve Challenges in Medical Imaging?
FPOAs are ideally suited to address many of these challenges in medical image processing. Operating deterministically at speeds to 1 GHz, with a rich array of compute and memory resources, MathStar FPOAs offer up to four times the performance of today’s top FPGAs for processing the filters, transforms, and searches that comprise the underlying algorithms.
Applicable MathStar libraries and IP cores
MathStar Application Assessments
| Assessment | Description | Document |
|
| CT Back Projection | Analysis of the fundamental Computed Tomography (CT) Back-Projection algorithm on the Arrix FPOA. Supports up to 1000 slices at 512x512 resolution. | ||
| Digital Beamforming | Analysis of multi-channel beamforming useful in ultrasound imaging and wireless communications | ||
| Scalable 3x3 Convolution Filter | Analysis of scalable 2D convolution filter block, scalable up to 15x15 on the Arrix FPOA. Symmetric or asymmetric. | ||
| Multi-tap FIR Filter | Analysis of 1 Giga sample per second Scalable FIR filter from 8 taps up to 128 taps, 16 bit precision. | ||
| Smith-Waterman |
|
If you would like to learn more about MathStar's medical imaging applications, please contact your local sales rep today.
Thank You!
