JPH0152780B2 - - Google Patents

Description

[Detailed Description of the Invention] [Field of Industrial Application] The present invention is an apparatus for analyzing red blood cells, which are components in blood, and more specifically, an analyzer for analyzing reticulocytes, which is considered to be particularly important in understanding the metabolic functions of red blood cells. The present invention relates to a blood analysis method and device for easily determining the proportion of red blood cells.

[Conventional technology]

Normally, when compared to mature red blood cells, reticulocytes are
They have a mesh pattern inside, and in a normal person, they are present at a rate of a few per 1,000 red blood cells, and are so-called red blood cells that have not fully matured and have been ejected as young. Knowing this number is extremely meaningful clinically.

Conventionally, a blood smear is stained with new methylene blue, etc., the image of the smear is captured into memory using a microscope camera, image processing is performed on the captured image information to separate areas of red blood cells, and pattern recognition is then performed. The method was used to determine whether the cells were reticulocytes or not.

[Problem to be solved by the invention]

The above method requires writing to memory, reading, and determining the area of red blood cells, all according to computer algorithms, and takes a considerable amount of time to determine whether or not it is a reticulocyte. ,
The disadvantage is that the memory capacity must be increased, resulting in a large-scale device similar to conventional leukocyte classification devices.

The present invention was made in order to eliminate the above-mentioned drawbacks, and provides a blood analysis method that is faster, reduces computer processing, and miniaturizes the apparatus by converting everything that used to be processed using computer software into hardware. The object of the present invention is to provide a device for the same.

[Means to solve the problem]

In order to achieve the above object, the blood analysis method of the present invention has the following (a):
(a) obtaining the number of red blood cells by dividing the total area above the cell level among the image signals counted by the area counting circuit 9 by the area of the red blood cells; (b) ) Out of the image signals, the address of the signal corresponding to the nucleus or net, which has a higher density than the protoplasm, is written in the address memory 8, the address data of this address memory 8 is read out, and the address corresponding to the address data of the image memory 6 is written. (c) determining whether or not the address data exists continuously in one cell area; (d) obtaining the reticulocyte count by dividing the total area of the reticulocytes by the area of the red blood cells; (e) ( The method includes the step of obtaining the ratio of reticulocytes to total red blood cells from the number of red blood cells obtained in a) and the number of reticulocytes obtained in step (b).

Further, as shown in FIG. 1, the blood analyzer of the present invention includes an imaging device 1 that outputs an image signal corresponding to the density of each point of an image and a synchronization signal necessary for scanning the image signal; An address signal generation circuit 2 that is connected to the imaging device and generates an address signal indicating the position of the image based on the synchronization signal, and an AD that is connected to the imaging device and the address signal generation circuit and converts the image signal from analog to digital.
A conversion circuit 3, a first comparison circuit 4 connected to the imaging device and passing a signal corresponding to the plasma concentration of the blood cells among the image signals, and a first comparison circuit 4 connected to the imaging device and passing a signal corresponding to the plasma concentration of the blood cells among the image signals; a second comparison circuit 5 that passes signals corresponding to the image density, and an image memory that is connected to an AD conversion circuit and an address signal generation circuit and stores a logic signal representing density at a predetermined address sent from the address signal generation circuit. 6, a gate circuit 7 connected to the second comparison circuit and the address signal generation circuit, an address memory 8 connected to the gate circuit, and an area counting circuit 9 connected to the first comparison circuit and the address signal generation circuit. , an arithmetic device 10 connected to this area counting circuit, an image memory, and an address memory, and an output device 11 connected to this arithmetic device.

[Effect]

The imaging device 1 outputs an image signal corresponding to the density of each point of the image and a synchronization signal necessary for scanning the image signal. Based on this synchronization signal, an address signal indicating the position of the image is output. Generating circuit 2 generates the signal. The image signal, which uses the shading of each point of the image as information, is converted from analog to digital by the AD conversion circuit 3, and is also converted from analog to digital by the AD conversion circuit 3.
The signal is sent to two comparison circuits 4 and 5 having comparison voltages V ₁ and V ₂ shown in FIGS. The first comparison circuit 4 is a circuit that passes a signal corresponding to the plasma concentration of blood cells, and the second comparison circuit 5 is a circuit that passes a signal corresponding to the nucleus or net, which has a higher concentration than the protoplasm. When the input signal is larger than the comparison voltages V ₁ and V ₂ of the comparison circuits 4 and 5, the outputs of the comparison circuits 4 and 5 are 1,
If it is small, it is 0. The image signal sent to the AD conversion circuit 3 is encoded according to the size of the signal, and then sent to the image memory 6. , a logic signal representing the concentration is stored. On the other hand, the signal at each point in the image exceeding the comparison voltage V ₁ is
It is counted by an area counting circuit 9. This area is greater than the total cell level, including white blood cells, whole red blood cells, etc. in a smear. When the second comparison circuit 5 detects a net, nucleus, foreign matter other than blood cells, etc. having a level higher than the protoplasm of blood cells, it sends a gate open signal to the gate circuit 7, and stores the address at that time in the address memory 8. It is for writing, and addresses are stored sequentially.

As described above, after counting the cell area for one screen, writing to the image memory 6, and writing the addresses of pixels with a comparison voltage of _V2 or higher, the arithmetic unit 10 performs the following calculations according to the built-in program. I do. That is, first, by dividing the total cell area by the area per red blood cell (approximately 50 μ ² ), you can find out how many red blood cell-sized cells there are. Next, the addresses (address data) in the address memory 8 are read out one after another, and
The content of the address adjacent to the address corresponding to the address data is read, and it is determined whether the signal is above the cell level, that is, whether the signal is due to dust or the like. Subsequently, the high-density area shows a continuous high value (as shown in Figure 3, and such a signal is obtained from the white blood cell image), or it is uneven (as shown in Figure 2). (such a signal can be obtained from reticulocyte images), or if the area beyond the cytoplasmic level is
Calculations are performed according to a predetermined algorithm that simultaneously determines whether or not the size of red blood cells is about ^50μ2 . In this way, only the pixel addresses corresponding to the reticulocyte networks are left in the memory, and all remaining dust and addresses corresponding to the white blood cell nuclei are erased. After that, the number of cells containing the net is determined by calculation. For example, one method is as follows. Read the image memory corresponding to the red blood cell area around the pixel determined to be a mesh, transfer all addresses that are above the level of the protoplasm, and if the addresses of the pixels to be transferred are the same, transfer them so that they do not overlap. , obtain the number of pixels corresponding to all addresses. Dividing this number of pixels by the number of pixels corresponding to the red blood cell area yields the number of reticulocytes.

As described above, the number of total cells and the number of total reticulocytes are determined by (number of pixels)/(number of pixels equal to the area of red blood cells). Normally, the number of reticulocytes or white blood cells relative to red blood cells is so small that it can be ignored except in cases of special diseases. Therefore,
Even if the total cell count obtained as described above is used as the red blood cell count, it will almost always fall within the margin of error. In this way, each image can be obtained for each part of the smear and analyzed to obtain the reticulocyte count when the number of red blood cells reaches 1000 or several thousand. Reticulocytes are usually expressed as a number per 1000 red blood cells. The absolute value of the reticulocyte count can be obtained from the ratio by inputting the separately determined number of red blood cells (cells/mm ³ ). This value is displayed or printed on the output device 11.

〔Example〕

Embodiments of the present invention will be described below based on the drawings. FIG. 1 shows an example of the configuration of the blood analyzer of the present invention, and FIGS. 2 and 3 are illustrations of the principle. The blood analyzer of the present invention includes an imaging device 1 such as a television camera, an address signal generation circuit 2 connected to the imaging device 1, and an AD conversion circuit 3 connected to the imaging device 1 and the address signal generation circuit 2. , a first comparison circuit 4 and a second comparison circuit 5 connected to the imaging device 1, an image memory 6 connected to the AD conversion circuit 3 and the address signal generation circuit 2, and a second comparison circuit 5 and the address signal generation circuit 2. A gate circuit 7 connected to the circuit 2, an address memory 8 connected to the gate circuit 7, an area counting circuit 9 connected to the first comparison circuit 4 and the address signal generation circuit 2, and this area counting circuit 9. , an arithmetic device 10 connected to the image memory 6 and address memory 8, and an output device 11 connected to this arithmetic device.
It includes.

In the apparatus configured as described above, the imaging device 1 outputs an image signal a corresponding to the shading of each point of the image and a synchronization signal b necessary for scanning the image signal. Therefore, the address signal generation circuit 2 generates an address signal indicating the position of the image based on the synchronization signal b. The image signal a, which uses the shading of each point of the image as information, is converted from analog to digital in the AD conversion circuit 3, and is converted into two signals having comparison voltages V ₁ and V ₂ as shown in FIGS. 2 and 3. The signal is sent to comparison circuits 4 and 5. Note that FIG. 2 shows the case of reticulocytes, and FIG. 3 shows the case of white blood cells. The first comparison circuit 4 is a circuit that passes a signal corresponding to the plasma concentration of blood cells, and the second comparison circuit 5 is a circuit that passes a signal corresponding to the nucleus or net, which has a higher concentration than the protoplasm. When the input signal is larger than the comparison voltages V ₁ and V ₂ of the comparison circuits 4 and 5, the outputs of the comparison circuits 4 and 5 are 1,
If it is small, it is 0. The image signal a sent to the AD conversion circuit 3 is encoded according to the size of the signal and then sent to the image memory 6, and is sent to a predetermined address in the image memory 6 sent from the address signal generation circuit 2. A logic signal representing the concentration is stored in the memory.
Note that the AD conversion is performed by the address signal generation circuit 2 using the sampling signal d corresponding to each point of the image when generating the address signal c based on the synchronization signal. On the other hand, signals at each point of the image exceeding the comparison voltage V ₁ are counted by the area counting circuit 9. This area is larger than the level of all the cells in the smear, including white blood cells, all red blood cells, etc. As shown in Figure 4, after the first row of scanning in the x direction is completed, it advances one step in the y direction. Scanning in the x direction is performed again. Each point is called a pixel, and one sampling signal d is obtained for each pixel, so the area counting circuit 9
Now, instead of counting the output of the first comparison circuit 4,
The output of the comparison circuit 4 may be used as a gate signal, and the sampling signal d may be counted. In the address signal generation circuit 2, as shown in FIG.
A sampling signal is generated based on the synchronization signal,
In general, the address in the y-axis direction is the synchronization signal,
The axial address indicates the count value resulting from counting the sampling signal using a logical symbol. The second comparison circuit 5 is a network having a level higher than the protoplasm of blood cells;
This is for sending a signal to open the gate to the gate circuit 7 when detecting dust other than nuclei or blood cells, and writing the address at that time into the address memory 8, where the addresses are sequentially stored.

As described above, after counting the cell area for one screen, writing to the image memory 6, and writing the addresses of pixels with a comparison voltage of _V2 or higher, the arithmetic unit 10 performs the following calculations according to the built-in program. I do. That is, first, by dividing the total cell area by the area per red blood cell (approximately 50 μ ² ), you can find out how many red blood cell-sized cells there are. Next, the addresses (address data) in the address memory 8 are read out one after another, and
The content of the address adjacent to the address corresponding to the address data is read, and it is determined whether the signal is above the cell level, that is, whether the signal is due to dust or the like. Subsequently, the high-density area shows a continuous high value (as shown in Figure 3, and such a signal is obtained from the white blood cell image), or it is uneven (as shown in Figure 2). (such a signal can be obtained from reticulocyte images), or if the area beyond the cytoplasmic level is
Calculations are performed according to a predetermined algorithm that simultaneously determines whether or not the size of red blood cells is about ^50μ2 . In this way, only the pixel addresses corresponding to the reticulocyte networks are left in the memory, and all remaining dust and addresses corresponding to the white blood cell nuclei are erased. After that, the number of cells containing the net is determined by calculation. For example, one method is as follows. Read the image memory corresponding to the red blood cell area around the pixel determined to be a mesh, transfer all addresses that are above the level of the protoplasm, and if the addresses of the pixels to be transferred are the same, transfer them so that they do not overlap. , obtain the number of pixels corresponding to all addresses (details will be explained later). Dividing this number of pixels by the number of pixels corresponding to the red blood cell area yields the number of reticulocytes. In order to remove white blood cells and dust or obtain the number of reticulocytes, it is also possible to use a conventional pattern recognition algorithm instead of relying on the above determination. However, commonly used pattern recognition algorithms require complicated calculation operations, require large-scale equipment, and may take several times as long to determine pixel connectivity or determine shape. be.

As described above, the number of total cells and the number of total reticulocytes are determined by (number of pixels)/(number of pixels equal to the area of red blood cells). Normally, the number of reticulocytes or white blood cells relative to red blood cells is so small that it can be ignored except in cases of special diseases. Therefore,
Even if the total cell count obtained as described above is used as the red blood cell count, it will almost always fall within the margin of error. In this way, each image can be obtained for each part of the smear and analyzed to obtain the reticulocyte count when the number of red blood cells reaches 1000 or several thousand. Reticulocytes are usually expressed as a number per 1000 red blood cells.

That is, the reticulocyte ratio typically
It is expressed as the number of reticulocytes per 1000 cells. Therefore, in order to calculate the reticulocyte ratio, for example, each time one screen of images is processed, the number of red blood cells and the number of reticulocytes are cumulatively added, and when the number of red blood cells exceeds a predetermined value, Reticulocyte count]/[Total red blood cell count] x 1000.

Furthermore, the absolute value of the reticulocyte count can be obtained from the ratio by inputting the separately determined number of red blood cells (cells/mm ³ ) (for example, using a hemocytometer). This value is displayed or printed on the output device 11.

In carrying out the present invention, strictly speaking, there is a problem if the area counted by the area counting circuit 9 is simply divided by the red blood cell area. This is because white blood cells and garbage other than red blood cells are also counted.

However, as mentioned above, there is a large difference between the number of red blood cells and the number of other things, so this can be considered to be within the margin of error and can be ignored.
And by ignoring it, processing becomes easier.

Further, the address memory 8 stores as data the address of a signal having a higher density than the protoplasm (a signal greater than the comparison voltage _V2 ) among the image signals. Beyond this V ₂ are garbage, reticulocyte meshwork, and white blood cell nuclei. That is, the address memory 8 stores the positions of dirt, nets, and kernels.

On the other hand, in the image memory 6, a logical symbol (density value) representing the magnitude of the signal, that is, the density, is stored at an address corresponding to the address of the image signal.

Then, reticulocytes and other particles are distinguished based on the following characteristic points.

(a) Reticulocytes have reticulum, but red blood cells do not.

(b) Garbage does not have surrounding protoplasm like blood cells.

(c) White blood cells have large nuclei, but reticulocytes have small networks, and there may be more than one within a single cell.

(d) White blood cells are larger than red blood cells, and reticulocytes are similar in size to red blood cells.

First, addresses (address data) in the address memory 8 are read one after another. Red blood cell data is not stored here. Address memory 8 contains garbage, nets, etc.
Nucleus address data is stored.

Next, the content (concentration value) of the address adjacent to the address corresponding to the above address data in the image memory 6 is read, and it is determined that the content (concentration value) of the address adjacent to the address corresponding to the above address data is read, and it is determined that the content (concentration value) is at the cell level (concentration corresponding to the comparison voltage V ₁ or concentration of protoplasm). Determine whether it is above. In other words, if it's not above the cellular level, it's garbage; if it's above the cellular level, it's reticulocytes or white blood cells. In addition, it is necessary to determine whether the high concentration area (the area with a concentration higher than the comparison voltage V ₂ ) shows a continuous high value (in the case of white blood cells) or whether it is in an uneven state (in the case of reticulocytes). Reticulocytes and white blood cells are distinguished by determining whether the area where the concentration value exceeds the cell level is comparable to that of red blood cells.

In the manner described above, it is determined that the address data in the image memory 6 is not dust or nucleus, but reticulum, and reticulocytes are identified.

Next, a process for determining the number of reticulocytes is performed using, for example, one of the methods described above.

As mentioned above, one method is to read the image memory corresponding to the red blood cell area around the pixel determined to be a mesh, transfer all addresses above the level of the protoplasm, and ensure that the addresses of the transferred pixels are the same. In the case of addresses, they are transferred so as not to overlap, and the number of pixels corresponding to all addresses is obtained.

That is, as shown in FIGS. 6 and 7,
The address of the portion where the reticulocyte and the area corresponding to the area of the red blood cell around the pixel of the reticulum overlap (shaded area) is transferred, and the number of pixels is determined. However, the portions that do not overlap (hatched portions) are not transferred. Such a transfer omission causes an error. but,
This situation rarely occurs and can be ignored. As shown in FIG. 8, this transfer leakage can be reduced by setting the area around pixels determined to be meshes to be larger than the area of red blood cells. Next, reticulocytes are determined by the method described above, that is, by dividing the obtained number of pixels by the number of pixels corresponding to the red blood cell area.

〔Effect of the invention〕

Since the blood analysis method and device of the present invention are configured as described above, it is possible to obtain the reticulocyte count from the area ratio without using complicated techniques such as conventional pattern recognition. It has the advantage of being able to be miniaturized and capable of high-speed analysis.

[Brief explanation of drawings]

Fig. 1 is a systematic explanatory diagram showing one embodiment of the blood analyzer of the present invention, Figs. 2 and 3 are explanatory diagrams of the principle, Fig. 2 shows the case of reticulocytes, and Fig. 3 shows the case of white blood cells. The case is shown below. FIG. 4 is an explanatory diagram showing a scanning state, FIG. 5 is a waveform diagram of a synchronization signal and a sampling signal, and FIGS. 6 to 8 are explanatory diagrams showing an example of a method for determining the number of reticulocytes. 1... Imaging device, 2... Address signal generation circuit, 3... AD conversion circuit, 4... First comparison circuit,
5... Second comparison circuit, 6... Image memory, 7...
Gate circuit, 8... Address memory, 9... Area counting circuit, 10... Arithmetic device, 11... Output device.

Claims (1)

[Claims] 1. The following steps (a) to (e): (a) The total area of the image signals counted by the area counting circuit 9, which is equal to or higher than the cell level, is calculated by the area of red blood cells. Step of obtaining the number of red blood cells by dividing (b) Writing the address of the signal corresponding to the nucleus or net, which has a higher concentration than the protoplasm, in the address memory 8, reading out the address data of this address memory 8, and converting the image into the image signal. (c) identifying garbage based on whether the value of an address adjacent to the address corresponding to the address data in the memory 6 is equal to or higher than the cell level; (c) if the address data is continuous in one cell area; (d) by dividing the total area of reticulocytes by the area of red blood cells; (e) obtaining the proportion of reticulocytes to total red blood cells from the number of red blood cells obtained in (a) and the number of reticulocytes obtained in (b); A blood analysis method characterized by: 2. An imaging device 1 that outputs an image signal corresponding to the shading of each point of the image and a synchronization signal necessary for scanning the image signal, and an imaging device connected to this imaging device that indicates the position of the image based on the synchronization signal. An address signal generation circuit 2 that generates an address signal; an AD conversion circuit 3 that is connected to the imaging device and the address signal generation circuit and converts the image signal from analog to digital; the first through which a signal corresponding to the concentration is passed;
A comparison circuit 4, a second comparison circuit 5 which is connected to the imaging device and passes signals corresponding to nuclei and nets having a higher concentration than the protoplasm out of the image signal, and a second comparison circuit 5 which is connected to the AD conversion circuit and the address signal generation circuit and which outputs the address signal. An image memory 6 that stores a logic signal representing the density at a predetermined address sent from the signal generation circuit.
, a gate circuit 7 connected to the second comparison circuit and the address signal generation circuit, an address memory 8 connected to the gate circuit, and an area counting circuit 9 connected to the first comparison circuit and the address signal generation circuit.
A blood analysis device comprising: a calculation device 10 connected to the area counting circuit, the image memory and the address memory; and an output device 11 connected to the calculation device.