X-ray digitization: medical/dental design considerations

Digital x-ray imaging is to bring about changes in diagnostic radiology.

In conventional x-ray systems, each component signals recession consumes more than 60% of the original x-ray signal energy. Each level in the system, x-ray signal will decline a certain amount, even for applications optimized for the stand-alone component is no exception. Therefore, only less than 40% of the original image information can be used to generate the image. By digital x-ray imaging to add digital detector, captured more than 80% of the original image information and use various tools to further improve the picture quality post-processing becomes possible. X-ray digitization of other benefits include: reduces the patient's medication dose, by exemption from photographic to reduce troubleshooting time, by eliminating the photographic chemicals to save costs, image data processing to enhance the area of concern and inhibit independent information; the image data and the other from RIS/HIS patient related information; network connection to any local rapid transmission of information; it only needs minimal space to put all the required information archive. X-ray digitization technology contains two different ways: direct conversion and indirect conversion.

Direct conversion

In the direct conversion, flat selenium x-ray detector directly absorbed and converted to independent pixel charge.

In indirect conversion, x-ray signal is first converted to a light signal, then converted to charge. Tile CCD (charge coupled device), uantitative array and uses indirect conversion technology. Tile CCD conversion technology with more than one CCD elements through optical fiber coupling to scintillator flat. Computer tomography by trapping optical stimulation plate, and through electrical signals to light produce image data. In two ways, and the pixel on the x-ray intensity proportional charge will be stored in thin-film transistor (TFT) storage capacity. A large number of the pixels are derived from flat panel detector (FDP). By reading electronic devices from reading the charge on the FDP and transform it into digital data.

The following block diagram shows through direct imaging to neutralize the FDP is converted to digital data readout electronics.

It has two chains: acquisition chain and bias chain. In the link, its front end for the analog front end that can be multiplexed different FDP (channel) on the storage capacitor charge and charge into voltage. Bias chain through media bias and door control circuit for TFT array generated bias voltage. Digital control and data adjustment by FPGA complete FPGA through high speed interface (serial interface, LVDS, optical interface) and an external image processing unit for high-speed serial communication. Temperature sensor, DAC, amplifiers and high input voltage tolerance capability of switching regulators are other key system block. Each block must have a startup PIN and synchronization frequencies to avoid interference with the acquisition of other blocks in the chain of crosstalk. FDP pixels will determine the number of ADC channels and the speed of the ADC. Static or dynamic collection also determining the speed of the ADC. Static acquisition mean to less than 1s time collecting single image, while the dynamic acquisition means that the image you want to refresh 30Hz speed. Dynamic acquisition for more detailed cardiovascular, fluorescent inspection or related applications, they are the same as the number of channels and need a faster data conversion. With 2MSPS and higher speed range and has excellent DC performance ADC can be good.

Indirect conversion

For indirect conversion, CCD output needs to be correlated double sampler (CDS).

Signal level of reduction of voltage and image signal level through the analog front end (AFE) into digital data. AFE's sampling rate from CCD array of pixels and a frame rate decision. In addition, the AFE will calibration sensor error, such as an undercurrent correction, offset voltage and defective pixels. Programmable gain amplifier (PGA) of existence, the PGA of linearity and gain range available is also very important, depending on the signal level. In the digital process, the digits will decide the contrast of the image. Usually, you need to set the initial data for accuracy than the ultimate digital image required digits high 2 to 4 bits of data. Therefore, if you need an 8-bit data of the final image, you first should be a 10-bit digitization to allow image processing rounding errors occur.

The main image quality metrics is a "quantum detection efficiency" (DQE), it combines contrast and SNR (signal to noise ratio), a percentage represents.

Contrast ratio is higher and the lower the noise, the higher the DQE. Contrast is the order of grayscale, it depends on the output resolution ADC; typically, 14-bit or 16-bit comparison suitable applications. SNR directs not only originate from ADC's SNR, and is the system of SNR, which is protected by x-ray dose, pixel size and impact of all electronic components. You can increase the x-ray dose and increase the photodiode spacing and electrical noise reduction to improve the SNR. Increased x-ray dose will have on patients or operator. Increase the photodiode space does not work, because doing so would decrease the spatial resolution. Decrease from the noise in the system for electronic devices will be the main challenges. System total noise is: all the noise on the signal chain of the square root of the composition (assumptions these noise component does not have an associated). This means including ADC, operational amplifier and benchmarks, all the parts must have a very low noise or severe filtering (if applicable). Temperature stability is another major challenge. The power consumption of the internal temperature rise may offset gray levels and make the image distortion, dynamic acquisition process. Therefore, the ADC, operational amplifier and benchmarks should have high temperature stability.