Components of a camera-computer system

Special line driver circuits are commonly used to drive the low-power scintillation camera signals to the computer. The use of line drivers not only ensures that the signals are not distorted but also protects the camera circuits from being damaged by the extra electronic load. The line drivers may also be used to alter the voltage levels of the signals so that they are of the

magnitude required by the computer interface. Most systems have

sample-and-hold circuits which retain the values of the position signals during the time that the computer is processing a detected event, even if the camera removes the signals from the line. Failures in these circuits may produce artefacts in the digital image, but usually will not affect the analogue operation of the camera. If the analogue and digital images differ, as illustrated in Fig. 7-1, these circuits should be considered as potential sources of the problem.

7.1.2.2. Analogue-to-digital conversion

The X and Y position signals must be converted to digital numbers to

be processed by the computer. There are several types of analogue-to digital

converters (ADC's) found in camera-computer systems. The most common

is the successive approximation converter, which makes sequential

estimates of the required numbers. Starting with the bit representing the

largest power of two, the converter sets the bit and then converts the binary

analogue image

digital image

15k c/s 56k c/s Count rate

Fig. 7-1. Effect of high count rate on digital image as compared with corresponding analogue image.

number to an analogue signal through a digital-to-analogue converter

(DAC). The amplitude of this analogue signal is compared with that of the

signal being converted (Fig. 7-2). If the signal being converted is smaller, the bit is turned off. If it is larger, the bit is left on. The ADC steps through each of the bits in the digital word, performing this process each time. For an eight-bit digital word (256 position values), the conversion takes eight cycles.

Signal from DAC

Signal being converted\

Bit

Fig. 7-2. Successive approximation analogue-to-digital conversion. The binary number corresponding to the analogue signal being converted is approximated bit by bit. At each step, the resulting analogue signal from a digital-to-analogue converter (DAC) is compared with the signal being converted.

7.1.2.3. Data processing

The data processor in the context of this document includes the CPU and the memory of the computer. The CPU in a conventional computer is the section that controls the timing and operation of the overall system. It also includes the arithmetic processing unit which performs the calculations and makes logical decisions. In newer computers the boundaries of the CPU are less clear, the single CPU being replaced by distributed microprocessors.

Although this is important to the system designer and to a certain extent to the user, it is not important for the understanding or execution of the quality control tests to be discussed.

The computer memory consists of a series of storage locations, or bins, into which data can be placed as words for later retrieval and manipulation.

Memory is characterized by the number of storage locations and the size of the individual word. The number of locations determines the amount of data and the size of programs that can be present at any given moment. The size of the memory word determines the magnitude of the number which can be stored at a given location as a binary number. Some word sizes have been given special names. The most common is the byte, which refers to a group of eight binary digits or bits.

In general, the size of the memory word determines the counts that can be collected in a digital image. Some computers allow the user to select the size that will be used for image collection. Use of an 8-bit storage element allows the collection of a count of 255 per image element (or pixel). Use of a 16 bit storage element accommodates numbers of up to 65 535 or ±<32 767 per pixel, depending on the particular computer. Computers may use other word sizes; at least one manufacturer uses a 10-bit storage element which limits the number per pixel to 1023.

The use of an 8-bit storage element for nuclear medicine imaging may represent a limitation and a potential source of error. In imaging procedures in which the radiopharmaceutical is concentrated in a small anatomical area, the pixels corresponding to this area quickly become saturated.

Depending on the particular computer, the computer may: 1) stop collecting, 2) continue collecting in the non-saturated areas while holding the saturated pixels at 255, thus severely distorting the quantitative data, or, 3) continue counting and allow the saturated pixel to "roll over" and lose multiples of 256 counts. Each of these may cause distortion of the quantitative data unless the system is capable of performing a suitable correction. It is important for the user to understand the clinical significance of such limitations and to choose the data collection mode appropriate to the clinical study to be performed.

7.1.2.4. Image formation

The output from the ADC is used in one of two ways by the computer

during data acquisition - list mode and frame mode. In list mode (Fig. 7-3),

the digital data representing the co-ordinates of photon interactions in the

crystal are simply stored as lists in memory analogous to those that a person

would record as numbers on a sheet of paper. In frame, or histogram, mode

(Fig. 7-4), the digital data are used to identify the address of a specific

memory location corresponding to the location of the interaction. The

contents of this memory location are then incremented by one. Frame-mode

Dans le document Quality Control of Nuclear Medicine Instruments 1991 | IAEA (Page 198-0)