JPH04316169A - Blood cell analyzing device - Google Patents

Abstract

PURPOSE:To improve the blood cell identifying and classifying accuracy of a blood cell analyzing device by correcting the criterion value of the device by introducing other inspection data, past personal data, and data of specimen preparing conditions. CONSTITUTION:Blood corpuscle number inspection data, such as the numbers of white corpuscles, red corpuscles, and blood platelets and hematocrit value, etc., biochemical inspection data, and other inspection data are respectively sent from a blood corpuscle inspecting device 3, biochemical inspecting device 4, and another inspecting device 6 to the blood cell analyzing device 1 through a host computer 5. In addition, past personal data are sent to the device 1 from the computer 5 and the smearing and staining conditions used at the time of preparing a specimen are sent to the device 1 from a specimen preparing device 2. The device 1 corrects a criterion value on the basis of the sent data. Then blood corpuscle identification is performed by executing a program adopting a corrected criterion value recognizing algorithm.

Description

[Detailed description of the invention]

0001

[Industrial Application Field] This invention recognizes blood cells, identifies and classifies them into mature cells and immature cells, and counts blood cells, making it suitable for detecting blood diseases such as leukemia and anemia. Regarding a cell analysis device.

0002

[Prior Art] In general, clinical tests include tests such as blood image analysis, blood cell counting, and biochemical tests, and various machines such as blood cell analyzers have been developed and used to automatically process these tests. There is. Additionally, systems have been developed and are in use that mutually and automatically transport tests used for testing between these machines.

0003

[Problems to be Solved by the Invention] The various conventional clinical devices described above are not configured to mutually exchange and use test results obtained by each device, so it is particularly difficult to transfer test data from other devices. In blood image analysis tests that often make recognition decisions based on the recognition, there is a problem in that the accuracy of the identification and classification results of this blood cell analyzer does not improve and is currently almost at its limit.

[0004] The present invention has been made in view of the above-mentioned problems, and provides a blood cell analyzer that is capable of importing data from other devices and is capable of obtaining more accurate identification and classification results. The purpose is to

[0005]

[Means and operations for solving the problem] The blood cell analyzer according to claim 1 of this application extracts characteristic parameters from a blood image and compares them with determination reference values using a predetermined recognition algorithm to identify and classify blood cells. and other testing equipment,
It is equipped with means for introducing test data, past personal data, and data on specimen preparation conditions from a host computer and a specimen preparation device, and means for correcting the judgment reference value using the introduced data.

[0006] This blood cell analyzer uses test data sent from other devices, past personal data, sample preparation conditions, and other data to determine the criteria for the algorithm that recognizes blood cells for each sample. is corrected to an appropriate value, and identification classification is performed based on this corrected judgment reference value. Therefore, the accuracy of identification and classification is greatly improved. The blood cell analyzer according to claim 2 of this application includes means for extracting feature parameters from a blood image, comparing them with determination reference values using a predetermined recognition algorithm, and identifying and classifying blood cells, and other testing equipment and a host. It is equipped with means for introducing test data, past personal data, and data on specimen preparation conditions from a computer and specimen preparation device, and means for correcting the recognition algorithm using the introduced data.

[0007] In this blood cell analyzer, the algorithm itself for recognizing blood cells is corrected and corrected based on data such as test data sent from other devices, past personal data, and specimen preparation conditions. Discrimination classification is performed based on recognition algorithms.

[0008]

[Examples] The present invention will be explained in more detail with reference to Examples below. FIG. 1 is a block diagram showing the configuration of an inspection system showing an embodiment of the present invention. This testing system is composed of a blood cell analyzer 1, a specimen preparation device 2, a blood cell counting testing device 3, a biochemical testing device 4, a host computer 5, and other testing devices 6. Data from the device 3, the biochemical testing device 4, and other testing devices 6 are transferred to the blood cell analyzer 1 via the host computer 5.

FIG. 2 is a block diagram showing the configuration of the blood cell analyzer 1. The blood cell analyzer 1 includes an automatic microscope 13 into which a specimen in a specimen cassette 11 is loaded and unloaded by an autofeeder 12, and an automatic A photographic imaging device 15 that captures the image of the microscope 13 via a half mirror 14a, a color TV camera 16 that also captures the image, and an A/D converter 17 that converts the captured image into a digital signal.
, an image memory 18 that stores digital image data, a feature extraction processor 19 that extracts feature parameters for recognition from the digital image, a recognition program 20 that performs identification and classification, and a , a recognition correction processor 21 that performs correction, a CPU 22 that performs various control operations, sensors 23 and 24 that detect adjustment signals, a WBC detection unit 25, an autofocus 26, a focus drive circuit 27, A stage drive circuit 28, an auto feeder controller 29,
Furthermore, it includes a VTR 30, an RGB monitor 31, a keyboard 32, and a printer 33. Also connected are an external tape puncher 7, a host computer 5, and other inspection devices 3 (4, 6).

In the above embodiment testing system, the blood cell counting testing device 3 collects blood cell counting testing data such as white blood cell count, red blood cell count, platelet count, hematocrit value, etc., the biochemical testing device 4 collects biochemical testing data, and other tests. Other test data is sent from the device 6 to the blood cell analyzer 1 via the host computer 5. Further, past personal data is sent from the host computer 5 to the blood cell analyzer 1. Furthermore, conditions for smearing and staining during specimen preparation are sent from the specimen preparation device 2 to the blood cell analyzer 1.

The blood cell analyzer 1 uses a recognition correction processor 11 based on data sent from other testing devices.
Correct the judgment reference value. As an example, if the recognition algorithm shown in FIG. 3 is used to identify blood cells from blood cell characteristic parameters, this recognition algorithm compares color tone, area, and nuclear shape as characteristic parameters. In step ST1, it is determined whether the color tone is greater than or equal to the criterion value α, in step ST2, it is determined whether the area is greater than or equal to the criterion value β, and in step ST3, it is determined whether or not the shape of the nucleus is greater than or equal to the criterion value γ. Of course, other characteristic parameters are used to discriminate blood cells, but three parameters are exemplified here. The judgment criteria α, β, and γ used here are standard values, but as mentioned above, when data is input from other inspection devices, the judgment criteria values α, β, and γ of these recognition algorithms change as shown in the figure. As shown in 4, α', β', and γ' are corrected. Then, a program that uses the corrected judgment reference value recognition algorithm is executed, and blood cell identification is performed. Since the corrected judgment reference values α', β', and γ' are values suitable for each test, a recognition algorithm is used for each specimen, and the accuracy of identification and classification is greatly improved.

[0012] In the above embodiment, only the judgment reference value used in the recognition algorithm is corrected using data from another device. The identification algorithm itself may be corrected. Alternatively, both the determination reference value and the recognition algorithm itself may be corrected, such as from what is shown in FIG. 3 to what is shown in FIG.

Furthermore, in the above embodiment, as shown in FIG. 1, all data other than the condition data at the time of specimen preparation are sent to the blood cell analyzer via the host computer 5, but this is shown in FIG. As shown, each data of the blood cell counting test device 3, biochemical test device 4, and other test devices 6 may be directly sent to the blood cell analyzer 1. Further, although not shown, all data including condition data during specimen preparation by the specimen preparation device 2 may be sent to the blood cell analyzer via the host computer.

Furthermore, in the embodiment shown in FIGS. 1 and 2 above,
Although a system for mutually and automatically transporting specimens used for testing is not added, a configuration in which a transport system is added may also be used (not shown). In addition, in the above embodiment, the recognition algorithm is configured by branching logic, but
A configuration using statistical methods such as discriminant functions is also possible. In this case, for example, as the discriminant function, the following formula

[0015]

[Math 1]

n: number of feature parameters P(i)
: Prior probability of occurrence of i species Σi : Covariance matrix (n×n) x of i species:
There is a minimum distance method that uses the feature parameter vector (1 x n) μi: the average value vector (1 x n) of the feature parameters of the i species to identify the i species with max{p(x|i)}. By correcting the average vector μi of the i types of feature parameters using data obtained from other inspection devices, a recognition algorithm suitable for each specimen is made, and the accuracy of identification and classification is improved.

[0017] Furthermore, the recognition algorithm may be constructed based on fuzzy theory. In this case, the membership function of the feature parameter is corrected using data obtained from another inspection device. This makes it possible to use a recognition algorithm that is suitable for each specimen and improve the accuracy of identification and classification. For example, examples of membership function correction are shown in FIGS. 8 and 9. FIG. 8 shows a standard membership function before correction of area, which is a feature parameter, and FIG. 9 shows a corrected membership function. Further, the fuzzy rule itself may be corrected using data obtained by another inspection device. Furthermore, both the membership function and the fuzzy rule may be corrected.

Furthermore, the present invention can be constructed by combining the techniques employed in each of the above-described embodiments.

[0019]

According to the present invention, a blood cell analyzer can receive test data, past personal data, and specimens from other test devices such as a blood cell counting test device and a biochemical test device, a host computer, a specimen preparation device, etc. By introducing data such as conditions at the time of production, and using this data to correct the criterion values of the algorithm for recognizing blood cells, or the algorithm itself, it is possible to significantly improve the accuracy of identification and classification.

[Brief explanation of drawings]

FIG. 1 is a block diagram of an inspection system showing an embodiment of the present invention.

FIG. 2 is a block diagram showing the configuration of a blood cell analyzer that constitutes the testing system.

FIG. 3 is a flow diagram showing a partial example of a recognition algorithm before correction of a judgment reference value in the same blood cell analyzer.

FIG. 4 is a flowchart showing a partial example of a recognition algorithm after correction of the determination reference value.

FIG. 5 is a flow diagram showing a partial example of another recognition algorithm after correction in the blood cell analyzer.

FIG. 6 is a flow diagram showing a partial example of still another recognition algorithm after correction in the blood cell analyzer.

FIG. 7 is a block diagram of an inspection system showing another embodiment of the invention.

FIG. 8 is a diagram showing an example of a membership function before correction when fuzzy inference is adopted as a recognition algorithm in a blood cell analyzer.

FIG. 9 is a diagram showing an example of a membership function after correction when the same fuzzy inference is adopted as a recognition algorithm.

[Explanation of symbols]

1 Blood cell analyzer 2 Specimen preparation device 3 Blood cell counting test device 4 Biochemical testing equipment 5 Host computer 6 Other inspection equipment 21 Recognition correction processor

Claims (2)

[Claims] 1. A means for extracting feature parameters from a blood image and comparing them with judgment reference values using a predetermined recognition algorithm to identify and classify blood cells; 1. A blood cell analyzer comprising: means for introducing personal data and data on specimen preparation conditions; and means for correcting the criterion value using the introduced data. 2. Means for extracting characteristic parameters from a blood image, comparing them with determination reference values using a predetermined recognition algorithm, and identifying and classifying blood cells; past personal data,
A blood cell analyzer comprising: means for introducing data on specimen preparation conditions; and means for correcting the recognition algorithm using the introduced data.