FPGA computing complex medical imaging equipment (1)

Medical imaging equipment in healthcare continues to play an increasingly important role, imaging technology is increasing and expanding new areas of application.

In order to meet market requirements, system design must have flexibility, also need to focus on improving the quality of diagnostic images, facilitate and reduce costs. In order to provide the expected functionality, and system developers start using FPGA.

Early diagnosis and treatment is promoting the use of imaging technology and hybrid technology integration, such as Positron Emission Tomography (PET) and computer tomography (CT).

To get the higher resolution image, the need for a precise geometric micro-array detector and complex software/hardware system on the photon and electron signal analysis. These systems must be on a growing number of image data (up to 10 Gbits/s) with high precision and very fast processing. In addition, in order to reduce the patient's medical expenses, equipment cost pressures still exist and, therefore, must have a long service life. Therefore need to be in its useful life upgrade features and algorithms more flexible systems. More and more engineers had to use programmable components, such as high-performance central processing unit (CPU) and field-programmable gate array (FPGA).

To develop efficient and flexible medical imaging equipment, must take into account the following factors:

• Imaging algorithm development requires advanced intuitive modeling tools for digital signal processing (DSP) for continuous improvement.

• Approximate real-time analysis of the performance requirements of system platforms at the same time equipped with software (CPU) and hardware FPGA.

• System architect and design engineers need in these platforms quickly split and debugging algorithm, using the latest tools and intellectual property (IP) library to accelerate their deployment and increase profitability.

Imaging algorithms

Image enhancement is often the convolution (linear) complete with filter.

High-pass filter to enhance image detail, but also makes noise more visible. Low-pass filter can suppress noise, but will make the image details. Most images contain content and exquisite part and part of the content of rough. Linear combinations of filtering technologies can enhance the former details and reduce the noise of the latter, by producing high-pass and low pass filter in the image and mask is a combination of both.

This technique works because the eye on the details area of noise is not sensitive.

Mask is a Sobel edge detection filter's output by smoothing. It uses to approximate (image) including details of parts and the use of zero to represent the details of the section is not included. High pass and low pass filter images by linear combinations are weighted by the mask, and noise reduction enhancement the details of the image.

Video picture stabilization and registration (VISAR) is a real-time video image dithering algorithm.

It developed to improve the video image quality, video data series for rotating and zooming effect, VISAR so that the image quality goes beyond simple horizontal and vertical image registration. VISAR by eliminating conversion, zoom and rotate to align the video image fields. Because VISAR allows the user to combine multiple video images, so that the noise is average to the frame. VISAR will also be extracted from the video of a static image of the jagged edge smoothing and image dithering correction-1/10 pixels.

VISAR algorithm can be used to:

• Will the cells under a microscope image clear

• Stability studies for retinal eye images

• Stable thermal infrared imaging

• In endoscopic surgery during steady camera and body movement

• View MRI video improvements in ultrasound technology for body movement to do revision

Wavelet transform is an analysis of algorithms, it overcomes the Fourier analysis of certain restrictions.

Fourier analysis in time domain signal from the transform to frequency domain so I lost time information. This is why when you see a signal's Fourier transform, it is not possible to tell you that a specific event occurs in any time. Many imaging signal contains important non-stationary or transient characteristics: in drift, trends, mutation, events, and/or end.

To help get event from signal information, Fourier transform is used for the analysis of a time only a small portion of the signal – called signals on Windows.

Recently, through the use of the band variable interval windowed on wavelet analysis be improved. Wavelet analysis allows a long interval to be more precise, low frequency information and shorter intervals to obtain high frequency information. Wavelet applications include non-contiguous and fault detection, self-similarity detection, signal suppression, signal and image noise, image compression and fast large matrix multiplication. Video and image processing (VIP) and DSP library for Wavelet operation provides a core standard components, including zoom, move, high/low-pass filtering, I/O decomposition and reconstruction.

Distributed vector processing is used to enable faster calculation algorithm.

S-transform (ST) combines the FFT and Wavelet transform features, revealing the frequency in space and time. Applications include texture analysis and noise filtering. However, the need for intensive computing ST, this makes the traditional CPU execution speed is too slow. But this can be through a combination of vector and parallel computing to solve, processing time compression 25 times. By FPGA implementation vector processors and parallel computing, and can greatly accelerate the computation-intensive algorithms.

1. as shown in the figure a common medical systems block diagram of a typical

Now, we will discuss promoting the FPGA integration to medical imaging equipment trends and core development results.

X

Optical imaging

From computer radiography (CR) digital radio (DR) changes are handled.

Digital flat panel detector in seconds on image processing, increase productivity, and does not require processing cassettes and films related chemicals.