JPS5880760A - Method and device for discrimination of reticular erythrocyte - Google Patents

Abstract

PURPOSE:To discriminate the reticular erythrocyte, by setting a window of a prescribed size around the coordinates of the detected picture element of the maximum value, producing a histogram from the concentration value and the concentration differential value among the picture elements within the window, and obtaining the exceeding frequencies over the threshold value of the histogram. CONSTITUTION:An A/D converter 2 works synchronously with the sampling pulse given from a sampling pulse generator 6 and gives an A/D conversion to the concentration of each picture element. In this case, the picture element is compared with the value of concentration to finally obtain the coordinates of the heighest concentration within a screen through a highest density detecting circuit 9. Then a window of a prescribed size is set the obtained coordinates, and the density value and the differential value of concentration are obtained by a subtractor circuit 19 for each picture element within the window. Thus a histogram is produced. Then the exceeding frequencies over the prescribed threshold value of the histogram is obtained and written to a concentration/ concentration differential value storing memory 16. In such way, the reticular erythrocyte is decided.

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention detects reticulocytes at high speed, which is considered to be of immediate importance in understanding the metabolic functions of red blood cells.

The invention also relates to a method and apparatus for effectively identifying the same.

In the case of a normal person with a mesh pattern in the area, 1000 red blood cells
These are young red blood cells that have not fully matured and have been ejected from the bloodstream at a ratio of a few to each individual. To have K
Become. 1. Or1. As mentioned above, the number of 14-shaped erythrocytes is only 1. It took a very long time to detect the cells because they were not condensed, and furthermore, the detected JI7+ cells were either reticuloid J/11 cells or red blood cells. Furthermore, conventional methods for determining whether a person is a deceased mountain slider have relied on pattern recognition, which requires a large-scale device and requires a considerable amount of time for measurement. In other words, if it takes time to identify the reticular blush, even though the number and proportion to the number of red blood cells is essential, a large amount of time will be lost, and the tail will be Since it is necessary to analyze several images and produce results from their total value, there is a disadvantage that the time required per sample is 4. The present invention has been made in order to solve the drawbacks described in F. - The coordinates of the pixel with the maximum value in each pixel of an image obtained with an imaging device such as a television camera are detected, and the coordinates of the pixel with the maximum value are detected in place of the coordinates. A histogram is created by setting a window of a predetermined size and determining the density value and density difference value of each pixel within one window.Then, the frequency of exceeding a predetermined threshold value of the histogram is determined, and the frequency or a predetermined value is calculated. The purpose of the present invention is to provide a method and apparatus that can quickly and effectively identify reticulocytes by determining that they are reticulocytes when they are detected.

The present invention will be explained below. FIG. 1 shows the scanning direction and concentration value signals of normal red blood cells and reticulocytes. In Figure 1, a reticulocyte has a higher concentration in the reticulum part than a normal red blood cell a, and the reticulum part is more protruding than a white blood cell, so the difference value in this part is a large value. is taking. That is, when scanning in the direction of the arrow, a signal with a flat trapezoidal concentration value is obtained for one arm of red blood cells a, and a signal with a protruding high concentration value and a large concentration difference value is obtained for one reticulocyte. It will be done.

Here, the density difference value is −Δ”= f (]jo + jo) f (as shown in the 21st z1-4-
io +, ko, () and -f(1o,,4
°) is the density value of the pixel at the coordinates (jo, jo). Staining the blood smear with pneumethylene tly- etc.1.
- When an image is captured by a microscope camera and scanned once, it is divided into each pixel, and each pixel of one image is compared and the coordinates of the pixel with the maximum density are determined. When each pixel is taken into the C memory. This coordinate is (l,
If j) may, -1 may be centered around this coordinate (", j') may, and -1 may ±" + j as shown in Figure 3.
Insert a window of max±In. In this case, it is preferable to set the value of m to a value of about 8 .mu.m so that one red blood cell can fit into the window when adjusted to the actual size as shown in FIG. The image corresponding to each pixel stored in the memory (one of the pixels within the window, i.e., i max
±m, , i Read density information of each pixel for 1naX Furukawa. Reading, +1 and 7 at the same time dark If (i, j
) vs. density difference value ΔfQ) is converted into a two-dimensional histogram. Suna-tcchi-You can't get the difference in intensity by comparing the density of a certain pixel with the pixel before that pixel. As shown in Figure 4 or Figure 5, the probes IL--L-- are read and converted one after another in a rough manner, and processing for one window is performed.

In the case of red blood cells in Tazunaku - In the case of a certain condition, there is no reticulum, and the difference value is smaller than that of reticulocytes, and in the case of one red blood cell, the histogram as shown in Section 41, Section 1. Because there is a part where the concentration of the white blood cells obtained is higher than that of the red blood cells, the distribution is toward the 10th corner of the graph, but it becomes a similar histogram. In the case of IA for reticulocyte red blood cells, both the concentration and the a degree difference value are large, as shown in FIG.

Therefore, as shown in Fig. 6, a predetermined 1 μm value i'1-△f1 is set for the density f and the density difference value Δf, and the frequency of the portion exceeding the threshold of 7P Shehis 1-gram is counted. It can be determined that it is a reticulocyte by examining the sum of F and -t. In the white surface f calculation, since it is generally biased upward, there are many areas with high concentration and low differential value, resulting in a form in which 1 -gram of His of 1 red blood cell is shifted upward. The determination of reticulocytes can be performed as described in -1- below.

An example of an apparatus for carrying out the above method is shown in FIG.

The reticulocyte identification device of the present invention uses a television camera which generates analog signals corresponding to the pixels and densities of five images, and the horizontal synchronization puffless signal and vertical synchronization pulse signal necessary to indicate the pixel position titf. ] The most image device wakes up 1 and-
An AD that converts analog signals from the imaging device 1 into AD.
Converter 2 and - A I] Transmit first memory data 1 to converter 2 via 1/distaro. The first memory data register 3 is connected to the first memory data register 3, which is connected to the first memory data register 3. Device 1〈Divide the horizontal synchronizing signal from tl into n equal parts, and divide the horizontal synchronization signal from tl into n parts corresponding to each pixel. L-less constant generator 6 Connected to Toco's sampling pulse generator B. Connected to the coordinate counting circuit 7 and the imaging device 1. J-coordinate counting circuit 8 for counting horizontal synchronization signals and AD converter 2
A maximum 1 seat detection circuit 9 is connected to the maximum density detection circuit 9 for detecting the maximum density in one screen, and a max register 10 is connected to the maximum density detection circuit 9 and the coordinate counting circuit 7 and stores the coordinates of the maximum 1 seat. , the maximum concentration detection circuit 111#; 9 and the J coordinate counting circuit 8 to store the j coordinate of the maximum concentration. j Max register 11 - Connected to sampling pulse generator 6 Control circuit 12 - Coordinate counting circuit 7 - j Coordinate counting circuit 8 1 Max register 10 - Connected to Max register 11 and image memory 5 The first memory address register 4, -i max register 10, a max register 11 . A window designation circuit 16 connected to the first memory address register 4 and the control circuit 12 and a computer 14 connected to the first control circuit 12
1 is connected to the first memory data register 1, and 1 is connected to the second memory address register 15 and 14 through the second memory address register 15.
The density difference value storage memo IJ 16 is connected to the density/density difference value storage memory 16 and the second memory data register 17 connected to the computer 14, and to the second memory data register 171C4'3. 7111 calculation NN
18 and a subtraction circuit 19 connected to the first memory take register 3 and the second memory address register 15 and calculating the difference value between the density value of the first memory take register 3 and the second memory address register 15.

In the device configured as above, each 7F pulse is divided into n equal parts so that the horizontal open IQ from the imaging device M1 is 111 times the 71st day! , the sample 11 y /) corresponding to each pixel is a pulse from the tv frequency generator 6: 4
1゛i,l-ru] The logic signal from the coordinate counting circuit 7 is
It is specified in the memory 7 address register 4. 1, which is the horizontal synchronizing signal.] The dramatic value of the coordinate counting circuit 8 corresponds to the value of J indicating which column is being scanned. Of course, the A/D converter 2 AD converts the degree of each pixel in synchronization with the sampling puff pulse from the sampler pulse generator 6. When each pixel is converted to AD data, the ratio of the density value to that of the pixel is calculated by 1 bit + 4LriilN': If the density is 1% lower, then - the value of the previous/manufactured value of 1 and its density. Coordinates of the pixel with 09,])max and input-h, depletion-
The maximum density/full output circuit 9 compares each pixel with the previous maximum value, and finally the coordinates (1°j) may of the pixel with the highest density on the screen are determined. -1maX -j max is stored in the comax register 10 and the Jmax register 11. Still, the control circuit 12 is activated by the final pulse from the sampling pulse generator 6, and the one window designation circuit 16 outputs imax±.
” - Set a window of jmax±m, and then directly pass through the first memory address register 4) - call each address within the window. The density information data is read out and sent to the memory 16 for storing 1 density/density difference values.At this time, the data indicating the 1 degree value is separated into 2 data itself, which becomes an address representing the density address. The histogram becomes the histogram, and the difference 11α from the previous concentration value is calculated by the subtraction circuit 19, and the two data are sent to the second memory address register 15 by the second memory address register 15, information on a single address combination, and 1.1,000 concentration/l agricultural product difference value. The data is written into the storage memory 16.

Therefore, this memory 16 is filled with the memorization of frequencies. Read the frequency at the intersection of the density address and the difference value address specified in the second register 15! , - Add 1 with the divine device 18 and write the new frequency in the original address again. Once the above processing is completed, the total is ', G'
The J-machine 14 reads fl and Δf as shown in FIG. When the sum of the frequencies of kneading, et al. exceeds a predetermined value λ, it is determined that it is a single-reticular blush.

As described above, according to the present invention, a predetermined window can be used regardless of pattern recognition.1. -Create a histogram for the cells within the window1. - In order to make judgments from histograms, Speed 1-' is fast and most of the calculations are done on hardware.'t1. - Only the final judgment and guidance signal commands are made by a computer, so the functionality of a small microcomputer is sufficient - Processing can be done at a considerable speed, making it smaller and lighter - This makes it possible to lower prices. In addition, there may be cases where two or more reticulocytes exist in one screen, and after removing the window portion, a window is designated around the maximum concentration point using the remaining image memory.1. Confirm that reticulocytes are present (1-1) and perform image replacement (1-). The ratio between the number of red blood cells and the number of reticulocytes can be calculated by finding the number of pixels in the trapezoidal area in one image memory, that is, the number of pixels above the plasma level in Figure 1, and dividing by the number of pixels equivalent to the area of one red sphere. The number is determined by dividing the reticulocyte count determined by the method described above by the total value per sample.

Generally, the reticulocyte count is calculated as a number per 1000 red blood cells.

[Brief explanation of drawings]

Fig. 1 is an explanatory diagram showing the scanning direction and concentration value signals of normal red blood cells and reticulocytes, Fig. 2 is an explanatory diagram showing the density difference values at pixels of the image of the imaging device, and Fig. 3 is an explanatory diagram showing the coordinates (
1, j) An explanatory diagram showing a state in which a window is provided around max. Figure 4 is a distribution diagram of the concentration of normal red blood cells - concentration difference value.
Figure 5 is a distribution diagram of reticulocyte concentration-concentration difference value - Figure 6
The figures are an explanatory diagram showing a state in which predetermined threshold values f and -Δf1 are set in FIG. FIG. DESCRIPTION OF SYMBOLS 1... Imaging device-2... AD converter, 6... First memory data register, 4... First memory address register, 5... Image memory 6... Sampling pulse generator -7...1 coordinate counting circuit, 8...J coordinate counting 1m path -9...maximum concentration detection circuit -101 max register, 11...j max register 12...control circuit, 16...-window designation circuit- 14... Calculator, 15.
・Second memory address register 16...Memory for storing concentration/concentration difference value 17...Second memory data register 18...Adder ~ 19...Subtraction circuit Patent applicant Toa Medical Electronics Co., Ltd. Agent Patent Attorney Shio Dema-7・2, l'2 ゛2. ' Fig. 4 Fig. ↓ E Δf Fig. 7

Claims (1)

[Claims] 1. Detecting the coordinates of the maximum pixel in each pixel of an image obtained by an imaging device such as a television camera, and setting a window of a predetermined size around the coordinates [2. A histogram is created by determining the density value and density difference value of each pixel, and the frequency at which the histogram exceeds a predetermined threshold is determined. A method for identifying reticulocytes, characterized in that when the frequency exceeds a predetermined value, it is determined to be reticulocytes. 2. An imaging device such as a television camera that generates an analog signal corresponding to the density of a pixel in an image and a horizontal synchronization puffless signal and a vertical synchronization pulse signal necessary to indicate the position of the pixel. An AD converter that performs AD conversion is connected to this AD converter via a first memory data register, and concentration information is written to a predetermined address specified by the first memory data register and the first memory address register. Divide the horizontal synchronizing signal from the image memory and one imaging device into n equal parts I2 Generate sampling pulses corresponding to each pixel 1. a sampling pulse generator for operating the AD converter; a one-coordinate counting circuit connected to the sampling pulse generator for counting one coordinate among the coordinates (i, j) of a pixel; and a horizontal synchronizing signal connected to the imaging device. a j-coordinate counting circuit for counting, and a maximum density detection circuit connected to the -AD converter and detecting the maximum density on the screen;
1 max register IL1 connected to this maximum concentration detection circuit and 1 coordinate counting circuit and storing 1 coordinate of maximum concentration.
．． ! =lit large concentration detection large sameness detection circuit jmax register which is connected to the coordinate counting circuit and stores the j coordinate of the maximum concentration and -sampling pulse W = 10H't K connected control circuit and -i coordinate counting circuit, J Coordinates tl-a circuit-
The first memory address register and the first memory address register are connected to the first memory address register and the image memory. The first memory address register and the first memory address register are connected to the first memory address register and the control circuit. a second memory address register connected to the first memory data register; a memory for storing concentration/concentration difference values connected to the computer via the second memory address register; and - the concentration/concentration difference value. a second memory data register connected to the storage memory and the computer; an adder connected to the second memory data register; and an adder connected to the first memory data register and the second memory address register; A reticulocyte identification device comprising a subtraction circuit for calculating a difference value from a concentration value.