WO2022021224A1 - 样本分类方法、装置和存储介质 - Google Patents

样本分类方法、装置和存储介质 Download PDF

Info

Publication number
WO2022021224A1
WO2022021224A1 PCT/CN2020/105878 CN2020105878W WO2022021224A1 WO 2022021224 A1 WO2022021224 A1 WO 2022021224A1 CN 2020105878 W CN2020105878 W CN 2020105878W WO 2022021224 A1 WO2022021224 A1 WO 2022021224A1
Authority
WO
WIPO (PCT)
Prior art keywords
classification
level
sample
classification results
results
Prior art date
Application number
PCT/CN2020/105878
Other languages
English (en)
French (fr)
Inventor
计得伟
李奔
周坤
曾凡顺
Original Assignee
深圳迈瑞生物医疗电子股份有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 深圳迈瑞生物医疗电子股份有限公司 filed Critical 深圳迈瑞生物医疗电子股份有限公司
Priority to CN202080102603.6A priority Critical patent/CN115867945A/zh
Priority to PCT/CN2020/105878 priority patent/WO2022021224A1/zh
Publication of WO2022021224A1 publication Critical patent/WO2022021224A1/zh

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition

Definitions

  • the present application relates to the technical field of sample classification of chip readers, and more particularly to a sample classification method, device and storage medium.
  • the current blood cell classification method usually scans, divides, and generates blood cell graphics automatically for the patient's blood samples prepared as blood smears by a chip reader, and gives the original classification results. Graphical reclassification.
  • a sample classification method includes: acquiring a sample image of a sample to be classified, displaying the sample image on a first display interface; receiving a sample image corresponding to the sample image to be classified A classification instruction for classifying samples; based on the classification instruction, the candidate classification results of the samples to be classified are displayed in the surrounding area of the sample image, so that the user can select the final classification results of the samples to be classified; The selection instruction of the candidate classification result is obtained, and the final classification result of the sample to be classified is obtained based on the selection instruction.
  • a sample classification apparatus includes a memory, a processor and a display, wherein the memory is used for storing a program and a sample image of a sample to be classified, the display is used for display under the control of the processor, the processor is configured to run a program in the memory to perform the following steps: acquiring a sample image of the sample to be classified, and displaying the sample on the first display interface by the display image; receive a classification instruction for classifying the to-be-classified sample corresponding to the sample image; based on the classification instruction, control to display the alternative classification result of the to-be-classified sample in the surrounding area of the sample image for the user Selecting the final classification result of the sample to be classified; receiving a selection instruction of the candidate classification result from the user, and obtaining the final classification result of the sample to be classified based on the selection instruction.
  • a storage medium is provided, and a computer program is stored on the storage medium, and the computer program executes the above-mentioned sample classification method when running.
  • the sample classification method, device, and storage medium display the candidate classification results of the samples to be classified in the surrounding area of the sample image of the samples to be classified, so that the user can quickly select the candidate classification results of the samples to be classified from the alternative classification results.
  • the final classification result improves the efficiency of sample classification, and the user operation is simple, reducing the possibility of classification errors.
  • FIG. 1 shows a schematic flowchart of a sample classification method according to an embodiment of the present application.
  • FIG. 2 shows an example of displaying alternative classification results for samples to be classified by the sample classification method according to an embodiment of the present application.
  • 3A and 3B illustrate another example in which the sample classification method according to the embodiment of the present application displays alternative classification results for the samples to be classified.
  • 4A and 4B illustrate yet another example of displaying alternative classification results for samples to be classified by the sample classification method according to an embodiment of the present application.
  • FIG. 5 shows a schematic block diagram of a sample classification apparatus according to an embodiment of the present application.
  • FIG. 6 shows an application example of the sample classification method according to an embodiment of the present application—a schematic structural diagram of a slicer.
  • FIG. 1 shows a schematic flowchart of a sample classification method 100 according to an embodiment of the present application.
  • the sample classification method 100 includes the following steps:
  • step S110 a sample image of the sample to be classified is acquired, and the sample image is displayed on the first display interface.
  • the sample image of the sample to be classified is displayed on the first display interface.
  • first display interface is so named to distinguish it from the "second display interface” that will appear later for displaying alternative classification results of samples to be classified, and has no other limiting meaning.
  • first display interface and the second display interface may be different display interfaces, or may be different states of the same display interface.
  • step S120 a classification instruction for classifying the to-be-classified sample corresponding to the sample image is received.
  • a classification instruction for classifying samples to be classified and inputted by a user through an input device can be received, such as dragging an image of the sample to be classified, performing a dragging instruction, long pressing commands, click commands, etc.
  • the dragging instruction may include: dragging the sample image of the sample to be classified by a preset distance, and the preset distance may be a small value, so as to avoid the inconvenience and error-prone problems caused by long-distance dragging.
  • alternative classification results of the samples to be classified can be obtained and displayed, as will be described in the following steps.
  • step S130 the candidate classification result of the sample to be classified is displayed in the peripheral area of the sample image based on the classification instruction, so that the user can select the final classification result of the sample to be classified.
  • an alternative classification result of the sample to be classified may be obtained first.
  • the candidate classification result of the sample to be classified may be obtained based on the automatic identification of the sample image of the sample to be classified.
  • the candidate classification results of the sample to be classified are displayed in the peripheral area of the sample image of the sample to be classified.
  • the peripheral area of the sample image can be understood as: taking the sample image as the center The area surrounding the sample image within the distance is exemplarily described below with reference to FIG. 2 .
  • FIG. 2 shows an example of displaying alternative classification results for samples to be classified by the sample classification method according to an embodiment of the present application.
  • the display interface includes respective sample images of a plurality of samples to be classified.
  • the alternative classification results are displayed in the surrounding area of the sample image S1. Includes: basophils, neutrophils, neutrophils, smear cells, and sediment.
  • basophils neutrophils, neutrophils, smear cells, and sediment.
  • the candidate classification result of the sample to be classified corresponding to the sample image S1 is displayed in a circular area centered on the sample image S1 and surrounding the sample image S1, it should be understood that this is only an example According to the teachings of the present application, displaying the alternative classification result of the sample to be classified in the surrounding area of the sample image of the sample to be classified does not have to be a circular area, as long as it is in the vicinity of the sample image, so that the user can conveniently, Choose quickly.
  • the method of displaying the alternative classification result of the sample to be classified in the surrounding area of the sample image of the sample to be classified can enable the user to quickly
  • the final classification result of the sample to be classified is selected from the alternative classification results, which improves the efficiency of sample classification, and the user operation is simple, reducing the possibility of classification errors.
  • the alternative classification results of the samples to be classified may be displayed on the second display interface.
  • the first display interface displaying the sample images of the samples to be classified and the display interface displaying the samples to be classified The second display interface of the alternative classification result may be a different display interface, or may be a different state of the same display interface.
  • the candidate classification results of the samples to be classified displayed in the peripheral area of the sample image may include at least one classification result obtained by classifying the samples to be classified corresponding to the sample image, and each candidate classification result may be Displayed in order based on their respective classification confidence.
  • the classification confidence of a classification result refers to the probability that the real classification result of the sample to be classified is the classification result. For example, after classifying a sample image of a white blood cell, one of the obtained alternative classification results is a neutrophil, and the classification confidence is 90%, which means that the probability of the white blood cell being a neutrophil is 90% .
  • the candidate classification results are sorted and displayed according to their respective classification confidences.
  • the classification confidence of a classification result is higher, it means that the classification result is more likely to be a real classification result. Therefore, based on the classification confidence
  • the way of sorting and displaying can make the user find the final classification result of the samples to be classified more quickly, and improve the classification efficiency.
  • the candidate classification results of the samples to be classified displayed in the peripheral area of the sample image may include at least two-level classification results, wherein the classification confidence of each of the previous-level classification results is higher than that of the latter-level classification results The classification confidence of each of the results can be displayed step by step based on a user instruction, so as to obtain a final classification result of the sample to be classified based on at least one classification result. This example is described below in conjunction with Figures 3A and 3B.
  • FIG. 3A and 3B show another example of displaying alternative classification results for samples to be classified by the sample classification method according to an embodiment of the present application, wherein FIG. 3A shows the first-level classification results of the samples to be classified, and FIG. 3B shows The second-level classification results of the samples to be classified are obtained.
  • the first-level classification results display the first classification results with the highest classification confidence, as shown in Figure 3A Eosinophilic, basophilic, neutrophilic, lymphatic, mononuclear; if the user believes that there is no real classification result of the sample to be classified in the first-level classification results or cannot determine whether there is a real classification of the sample to be classified in the first-level classification structure As a result, the pop-up of the second-level classification results can be triggered, and the second-level classification results show several classification results with the next highest classification confidence, as shown in Figure 3B, primitive, neutral late juvenile, atypical lymphoid, and young monocyte. It should be understood that this is only exemplary, and in practical applications, the displayed alternative classification results may include more than two-level classification results.
  • the user when the user triggers the display of the next-level classification results, the user can execute a hovering instruction, a long-pressing instruction, or a clicking instruction, etc. on any one of the previous-level classification results.
  • an "Other” option may be provided in the previous-level classification result, so that the user can trigger the next-level classification result through the "Other” option, as shown in the "Other” option in FIG. 3A .
  • the alternative classification result may further include an "unclear classification" option, as shown in FIG. 3B , this option may be used in a scenario where the user cannot determine the final classification result.
  • the hovering instruction includes hovering the pointer on the preset area of the selected object to select the selection method of the object.
  • the previous-level classification result may be closer to the sample image than the latter-level classification result, for example, the latter-level classification result may be displayed in the surrounding area of the former-level classification result, for example, as shown in FIG. 3B Yes, the second-level classification results are displayed in the surrounding area of the first-level classification results.
  • This way of display is clear and convenient for users to make more flexible choices, such as selecting from the second-level classification results and returning to the first-level classification results.
  • the previous-level classification result may not be displayed when the subsequent-level classification result is displayed, and an option for returning to the previous-level classification result may be provided at the same time.
  • the first-level classification results in the aforementioned at least two-level classification results can be displayed in at least a part of the area surrounded by the first circumference, and for the remaining level-level classification results in the at least two-level classification results, the latter level
  • the classification result is displayed in at least a part of the area surrounded by the second circumference around the previous classification result, the first circumference is concentric with the second circumference and the radius of the second circumference is larger than that of the first circumference, for example, as shown in FIG. 3B .
  • the classification results are displayed step by step in the form of a circle.
  • the circle shape has a high area utilization rate, is convenient for classification, and is visually neat, so that the user can visually select the final classification result of the sample to be classified, which is convenient, fast and accurate. choice.
  • the aforementioned example of hierarchically displaying the classification results based on the classification confidence allows the user to find the final classification result in some less candidate classification results without selecting the final classification result of the sample to be classified from a large number of options at one time. Further improve the classification efficiency.
  • the classification results with the highest classification confidence can be further displayed in the following manner to distinguish them from other classifications.
  • the display method of the classification results of the confidence level is displayed, or, the classification results of different classification confidence levels are displayed separately.
  • the distinguishing display includes at least one of the following: displaying the classification results of different classification confidences with icons of different colors; displaying the classification results of different classification confidences with icons of different sizes; The icons at different distances of the image show the classification results of different classification confidences; display the classification results of different classification confidences in the text of different attributes; display the classification results of different classification confidences in areas with different shapes; The numerical value of the degree is displayed on the corresponding classification result.
  • the classification result with the highest classification confidence is displayed in a display manner that is distinguished from the classification results of other classification confidences, or the classification results with different classification confidences are displayed in a different manner from each other, which can further strengthen the user's vision. As a result, the user can more clearly see the classification results with high classification confidence, which can not only further improve the classification efficiency, but also further reduce the possibility of classification errors due to such a differentiated display.
  • the alternative classification results of the samples to be classified displayed in step S130 may include at least two-level classification results, wherein the at least one-level classification results are divided into multiple groups, and the multiple groups include All possible classification results in this level of classification are listed, and the classification items included in any grouping in the previous level of classification results are the latter-level classification results, and the display of the alternative classification results includes: displaying each level by level based on user instructions. The first-level classification result is used to obtain the final classification result of the sample to be classified based on the at least one-level classification result.
  • the first-level classification results display the grouping results of the samples to be classified
  • the second-level classification results display the first-level classification results.
  • the subordinate result of any group in the classification result that is, the group or classification result further included in the group, and so on. This example is described below in conjunction with Figures 4A and 4B.
  • FIG. 4A and 4B show another example of displaying alternative classification results for samples to be classified by the sample classification method according to an embodiment of the present application, wherein FIG. 4A shows the first-level classification results of the samples to be classified, and FIG. 4B shows The second-level classification results of the samples to be classified are obtained.
  • the first-level classification results show the categories of groups to which the samples to be classified may belong, such as eosinophilic, basophilic, neutrophilic, mononuclear, and lymphatic as shown in Figure 4A; the user can Based on the first-level classification result, the group to which the sample to be classified belongs is selected as the final classification result of the to-be-classified sample, or the second-level classification result can be triggered to accurately select the next-level category of a certain group in the first-level classification result.
  • Figure 4B primitive, neutral late-juvenile, atypical lymphoid, juvenile mononucleus. It should be understood that this is only exemplary, and in practical applications, the displayed alternative classification results may include more than two-level classification results.
  • the user when triggering the display of the next-level classification results, the user can execute a hovering instruction, a long-pressing instruction, or a clicking instruction, etc. on any one of the previous-level classification results.
  • an "other" option can be provided in the previous classification result, so that the user can trigger the subsequent classification result through the "other" option, as shown in the "other group” option in FIG. 4A .
  • the alternative classification result may further include an option of "unknown classification", as shown in FIG. 4B , this option may be used in a scenario where the user cannot determine the final classification result.
  • the previous-level classification result may be closer to the sample image than the latter-level classification result, and the latter-level classification result may be displayed in the surrounding area of the previous-level classification result, for example, as shown in FIG. 4B , the second-level classification results are displayed in the surrounding area of the first-level classification results.
  • This display mode is clearly structured and facilitates more flexible choices for users, such as selecting from the second-level classification results and returning to the first-level classification results.
  • the first-level classification results in the aforementioned at least two-level classification results can be displayed in at least a part of the area surrounded by the first circumference, and for the remaining level-level classification results in the at least two-level classification results, the latter level
  • the classification result is displayed in at least a part of the area surrounded by the second circumference around the previous classification result.
  • the first circumference is concentric with the second circumference and the radius of the second circumference is larger than that of the first circumference, for example, as shown in FIG. 4B .
  • the classification results are displayed step by step in the form of a circle.
  • the circle shape has a high area utilization rate, is convenient for classification, and is visually neat, so that the user can visually select the final classification result of the sample to be classified, which is convenient, fast and accurate. choice.
  • each classification item included in each level of classification result in the at least two-level classification results in the alternative classification results described in conjunction with FIG. 4A and FIG. 4B may also be classified according to the level of classification confidence. differentiated display.
  • each classification item included in each level of classification result represents a possible grouping result or classification result of the sample to be classified, and the result also has a confidence level, and each classification item is distinguished from each other based on the confidence level and displayed. It can further strengthen the user's visual effect and improve the efficiency of the user's selection of classification results.
  • the foregoing examples of classification results based on classification confidence and/or grouping can improve the efficiency of user information screening. Only when the previous level does not provide the target classification result, the next level, or the previous The first level provides guidance information for the latter level, so that users do not need to screen a large amount of information at the same time, so that the classification efficiency can be further improved.
  • the classification instruction received in step S120 may further include an instruction to input a classification keyword; in this embodiment, the alternative classification results of the samples to be classified displayed in step S130 may include Classification results for keywords.
  • the corresponding candidate classification results are displayed according to the classification keywords input by the user, and the number of displayed candidate classification results can also be reduced, so that the user can quickly select the final classification results of the samples to be classified.
  • the alternative classification results of the samples to be classified displayed in step S130 may also include classification results obtained based on historical data, where the historical data includes final classification results of previous samples.
  • the user's selection of the final classification result of the previous sample is used as historical data to classify the current sample, so the current sample can be classified in combination with the user's historical operations, which can further improve the displayed alternative classification results It is the possibility of correct classification results (true classification results), thereby further improving the classification efficiency.
  • the display position of the classification results obtained based on historical data can be closer to the sample image of the sample to be classified;
  • the above classification results based on the classification confidence ranking are displayed in a display position closer to the sample image of the sample to be classified.
  • step S140 a user's selection instruction for the candidate classification result is received, and a final classification result of the sample to be classified is acquired based on the selection instruction.
  • the user can select the final classification results of the samples to be classified from them through a selection instruction.
  • the selection instruction may include a long-press instruction, a click instruction, and the like on any of the alternative classification results.
  • the final classification result of the sample to be classified can be obtained.
  • the final classification results of the samples to be classified can be displayed.
  • the method 100 may further include (not shown): after obtaining the final classification result of the current sample to be classified, performing the following operations on the remaining samples located in the same initial group as the current sample to be classified Any of: automatically display the alternative classification results of the remaining samples of the same initial grouping; generate the final classification results of the remaining samples of the same initial grouping based on the final classification results of the current samples to be classified; wherein, the initial grouping It is generated by grouping based on the initial classification result of the sample.
  • the current sample to be classified is the sample to be reclassified.
  • the reclassification is the verification or correction of the initial classification result, and the sample is obtained after reclassification.
  • the reclassification result is the final classification result.
  • the sample classification method 100 of the present application is used to reclassify the samples, that is, the samples have been initially classified before the reclassification, and multiple samples may be initially classified into the same group.
  • the alternative classification of other samples can be automatically started without inputting classification instructions for the sample images of other samples in the same group.
  • the candidate classification result of the sample to be classified is displayed in the surrounding area of the sample image of the sample to be classified, so that the user can quickly select the sample to be classified from the candidate classification results.
  • the final classification result improves the efficiency of sample classification, and the user operation is simple, reducing the possibility of classification errors.
  • the candidate classification results of the samples to be classified can also be displayed in a hierarchical manner, which can improve the efficiency of user information screening. level, or the former level provides guidance information to the latter level, so that the user does not need to screen a large amount of information at the same time, so that the classification efficiency can be further improved.
  • alternative classification results with different classification confidences of the samples to be classified can also be displayed separately, which can further strengthen the user's visual effect, so that the user can more clearly see the classification confidence.
  • the classification results with high degree of classification can not only further improve the classification efficiency, but also further reduce the possibility of classification errors due to such a differentiated display.
  • FIG. 5 shows a schematic block diagram of a sample classification apparatus 500 according to an embodiment of the present application.
  • the sample classification apparatus 500 may include a memory 510 , a processor 520 and a display 530 .
  • the memory 510 stores programs for implementing corresponding steps in the sample classification method 100 according to the embodiment of the present application.
  • the processor 520 is configured to run the program stored in the memory 510 to execute corresponding steps of the sample classification method 100 according to the embodiment of the present application, and the display 530 is configured to display information based on the control of the processor 520 .
  • the display 530 is configured to display information based on the control of the processor 520 .
  • the processor 520 when the computer program is run by the processor 520, the processor 520 causes the processor 520 to perform the following steps: acquiring a sample image of the sample to be classified, and displaying it on the first display interface by the display 530 the sample image; receiving a classification instruction for classifying the to-be-classified sample corresponding to the sample image; controlling, based on the classification instruction, to display an alternative classification result of the to-be-classified sample in a peripheral area of the sample image, for the user to select the final classification result of the sample to be classified; receive the user's selection instruction on the candidate classification result, and obtain the final classification result of the sample to be classified based on the selection instruction.
  • the candidate classification result is obtained by recognizing the sample image by the processor 520 .
  • the candidate classification result includes at least one classification result obtained by classifying the sample to be classified corresponding to the sample image, and each of the candidate classification results is ranked based on the respective classification confidence show.
  • the ranking display includes: the higher the classification confidence, the closer the display position of the corresponding classification result is to the sample image.
  • the alternative classification results include at least two-level classification results, wherein the classification confidence of each of the previous-level classification results is higher than the classification confidence of each of the subsequent-level classification results
  • the sorting and displaying include: displaying each level of classification results step by step based on a user instruction, so as to obtain a final classification result of the samples to be classified based on at least one level of classification results.
  • the classification results with the highest classification confidence are distinguished from other classifications
  • the display method of the classification results of the confidence level is displayed, or, the classification results of different classification confidence levels are displayed separately.
  • the differentiated display includes at least one of the following: displaying classification results of different classification confidences with icons of different colors; displaying classification results of different classification confidences with icons of different sizes; Icons at different distances relative to the sample image display the classification results of different classification confidences; display the classification results of different classification confidences with text with different attributes; display the classification results of different classification confidences in areas with different shapes; The numerical value of the classification confidence of the classification result is displayed on the corresponding classification result.
  • the alternative classification results include at least two-level classification results, wherein the at least one-level classification results are divided into multiple groups, and the multiple groups include all possible classifications in this level of classification
  • the classification item included in any group in the previous-level classification result is the latter-level classification result
  • the display of the alternative classification results by the display 530 includes: displaying each level of classification results step by step based on user instructions, so as to use to obtain the final classification result of the sample to be classified based on the at least one-level classification result.
  • each classification item included in each level of classification result in the at least two-level classification result is displayed according to the level of classification confidence.
  • the latter-level classification result is displayed in a peripheral area of the former-level classification result.
  • the first-level classification results in the at least two-level classification results are displayed in at least a part of the area surrounded by the first circumference, and for the remaining level-level classification results in the at least two-level classification results,
  • the latter-level classification result is displayed in at least a partial area surrounded by a second circumference around the former-level classification result, the first circumference is concentric with the second circumference and the radius of the second circumference is greater than the radius of the second circumference.
  • the radius of the first circle is a part of the area surrounded by the first circumference
  • the former-level classification result is closer to the sample image than the latter-level classification result.
  • the alternative classification result is displayed on a second display interface, and the first display interface and the second display interface are different display interfaces.
  • the candidate classification result and the sample image are displayed in different states of the first display interface.
  • the classification instruction includes any one of the following instructions executed on the sample image: a drag instruction, a long press instruction, and a click instruction.
  • the classification instruction includes an instruction to input a classification keyword
  • the alternative classification results include classification results based on the classification keywords
  • the user instruction includes the classification instruction and any one of the following instructions executed on any one of the previous classification results: a hover instruction, a long press instruction, and a click instruction.
  • the display 530 is further configured to: display the final classification result of the sample to be classified.
  • the alternative classification results displayed for the samples to be classified further include classification results obtained based on historical data, where the historical data includes final classification results of previous samples.
  • the display position of the classification results obtained based on historical data is closer to the sample image than the classification results displayed based on the classification confidence ranking; or relative to the classification results obtained based on historical data
  • the classification results are sorted and displayed based on the classification confidence, and the display positions of the displayed classification results are closer to the sample image.
  • the processor 520 when the computer program is run by the processor 520, the processor 520 further causes the processor 520 to perform the following steps: after obtaining the final classification result of the sample to be classified, compare the sample to be classified with the sample to be classified The remaining samples in the same initial group perform any one of the following operations: automatically display the alternative classification results of the remaining samples in the same initial group on the display 530; The final classification results of the remaining samples in the initial grouping; wherein the initial grouping is generated based on the initial classification results of the samples, the samples to be classified are the samples to be reclassified, and the reclassification is the The verification or correction of the initial classification result, the reclassification result obtained after the sample is reclassified is the final classification result.
  • the peripheral area includes an area surrounding the sample image within a predetermined distance with the sample image as the center.
  • the alternative classification results include at least two-level classification results
  • the at least two-level classification results are displayed step by step based on user instructions
  • the final classification results of the samples to be classified are based on at least one-level classification results result obtained.
  • a storage medium in which program instructions are stored, and when the program instructions are executed by a computer or a processor, corresponding steps of the sample classification method of the embodiments of the present application are performed.
  • the storage medium may include, for example, a memory card for a smartphone, a storage unit for a tablet computer, a hard disk for a personal computer, a read only memory (ROM), an erasable programmable read only memory (EPROM), a portable compact disk read only memory (CD). - ROM), USB memory, or any combination of the above storage media.
  • a computer-readable storage medium can be any combination of one or more computer-readable storage media.
  • a computer program is also provided, and the computer program can be stored in the cloud or on a local storage medium.
  • the computer program is run by a computer or a processor, it is used to execute the corresponding steps of the sample classification method of the embodiments of the present application.
  • Figure 6 shows a schematic diagram of the structure of the chip reader.
  • the reading machine is a kind of medical equipment, and the reading machine includes an objective lens 601 and a reading platform 602 arranged oppositely, wherein a glass slide 603 is placed on the reading platform 602, and the glass slide 603 is dripped with
  • the objective lens 601 is connected to the camera, so that the camera can photograph the sample on the glass slide 603 through the objective lens 601 to obtain a sample image.
  • the reading machine outputs the initial classification result of the sample corresponding to the sample image on the human-computer interaction device (not shown) by automatically recognizing the sample image.
  • the sample classification method according to the embodiment of the present application can be used For reclassifying the initially classified samples, the user can invoke the sample classification method provided by the embodiment of the present application by operating the human-computer interaction device, and quickly select the sample classification results displayed in the surrounding area of the sample image from the alternative classification results. Efficiently select the final classification result of the samples, as described above.
  • the sample classification method, device, and storage medium display the candidate classification result of the sample to be classified in the surrounding area of the sample image of the sample to be classified, so that the user can quickly select the classification result from the candidate classification result. Select the final classification result of the sample to be classified, improve the efficiency of sample classification, and the user operation is simple, reducing the possibility of classification errors.
  • the sample classification method, device, and storage medium can also display the alternative classification results of the samples to be classified in a hierarchical manner, which can improve the efficiency of user information screening.
  • Various component embodiments of the present application may be implemented in hardware, or in software modules running on one or more processors, or in a combination thereof.
  • a microprocessor or a digital signal processor (DSP) may be used in practice to implement some or all functions of some modules according to the embodiments of the present application.
  • DSP digital signal processor
  • the present application can also be implemented as a program of apparatus (eg, computer programs and computer program products) for performing part or all of the methods described herein.
  • Such a program implementing the present application may be stored on a computer-readable medium, or may be in the form of one or more signals. Such signals may be downloaded from Internet sites, or provided on carrier signals, or in any other form.

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Data Mining & Analysis (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Evolutionary Biology (AREA)
  • Evolutionary Computation (AREA)
  • Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Artificial Intelligence (AREA)
  • Information Retrieval, Db Structures And Fs Structures Therefor (AREA)

Abstract

一种样本分类方法（100）、装置（500）和存储介质，样本分类方法（100）包括：获取待分类样本的样本图像（S1），在第一显示界面上显示样本图像（S1，S110）；接收用于对样本图像（S1）对应的待分类样本进行分类的分类指令（S120）；基于分类指令在样本图像（S1）的周边区域显示待分类样本的备选分类结果，以供用户选择待分类样本的最终分类结果（S130）；接收用户对备选分类结果的选择指令，并基于选择指令获取待分类样本的最终分类结果（S140）。将待分类样本的备选分类结果显示在待分类样本的样本图像（S1）的周边区域，使得用户可以快速从备选分类结果中选择待分类样本的最终分类结果，提高样本分类的效率，且用户操作简单，减少发生分类错误的可能性。

Description

样本分类方法、装置和存储介质
说明书
技术领域
本申请涉及阅片机样本分类技术领域，更具体地涉及一种样本分类方法、装置和存储介质。
背景技术
目前的血细胞分类方法通常是由阅片机针对被制备为血涂片的病人血液样本自动完成血细胞图形扫描、分割、生成血细胞图形，并给出原始分类结果，在此基础上人工进行审核和血细胞图形的重分类。
目前在进行血细胞重分类时，用户主要通过以下方式来进行：1)拖动细胞图片到参数表格中对应的参数行；2)拖动细胞图片到目标分类的图片区域；3)右键细胞图片弹出下拉框，在下拉框中选取了目标分类。上述的前两种方式拖动距离较远，操作不便，且第一种方式表格参数行多，容易拖动到错误分类中；上述的第三种方式在下拉框中显示所有分类项，无法快速查找和选中目标类型，效率低下且易发生分类错误。
发明内容
为了解决上述问题中的至少一个而提出了本申请。根据本申请一方面，提供了一种样本分类方法，该方法包括：获取待分类样本的样本图像，在第一显示界面上显示所述样本图像；接收用于对所述样本图像对应的待分类样本进行分类的分类指令；基于所述分类指令在所述样本图像的周边区域显示所述待分类样本的备选分类结果，以供用户选择所述待分类样本的最终分类结果；接收用户对所述备选分类结果的选择指令，并基于所述选择指令获取所述待分类样本的最终分类结果。
根据本申请另一方面，提供了一种样本分类装置，所述装置包括存储器、处理器和显示器，其中，所述存储器用于存储程序和待分类样本的样本图像，所述显示器用于基于所述处理器的控制进行显示，所述处理器用 于运行所述存储器中的程序以执行以下步骤：获取所述待分类样本的样本图像，并由所述显示器在第一显示界面上显示所述样本图像；接收用于对所述样本图像对应的待分类样本进行分类的分类指令；基于所述分类指令控制在所述样本图像的周边区域显示所述待分类样本的备选分类结果，以供用户选择所述待分类样本的最终分类结果；接收用户对所述备选分类结果的选择指令，并基于所述选择指令获取所述待分类样本的最终分类结果。
根据本申请再一方面，提供了一种存储介质，该存储介质上存储有计算机程序，计算机程序在运行时执行上述样本分类方法。
根据本申请实施例的样本分类方法、装置和存储介质将待分类样本的备选分类结果显示在待分类样本的样本图像的周边区域，使得用户可以快速从备选分类结果中选择待分类样本的最终分类结果，提高样本分类的效率，且用户操作简单，减少发生分类错误的可能性。
附图说明
图1示出根据本申请实施例的样本分类方法的示意性流程图。
图2示出根据本申请实施例的样本分类方法针对待分类样本显示备选分类结果的一个示例。
图3A和图3B示出根据本申请实施例的样本分类方法针对待分类样本显示备选分类结果的另一个示例。
图4A和图4B示出根据本申请实施例的样本分类方法针对待分类样本显示备选分类结果的再一个示例。
图5示出根据本申请实施例的样本分类装置的示意性框图。
图6示出根据本申请实施例的样本分类方法的应用示例——阅片机的结构示意图。
具体实施方式
为了使得本申请的目的、技术方案和优点更为明显，下面将参照附图详细描述根据本申请的示例实施例。显然，所描述的实施例仅仅是本申请的一部分实施例，而不是本申请的全部实施例，应理解，本申请不受这里描述的示例实施例的限制。基于本申请中描述的本申请实施例，本领域技 术人员在没有付出创造性劳动的情况下所得到的所有其他实施例都应落入本申请的保护范围之内。
在下文的描述中，给出了大量具体的细节以便提供对本申请更为彻底的理解。然而，对于本领域技术人员而言显而易见的是，本申请可以无需一个或多个这些细节而得以实施。在其他的例子中，为了避免与本申请发生混淆，对于本领域公知的一些技术特征未进行描述。
应当理解的是，本申请能够以不同形式实施，而不应当解释为局限于这里提出的实施例。相反地，提供这些实施例将使公开彻底和完全，并且将本申请的范围完全地传递给本领域技术人员。
在此使用的术语的目的仅在于描述具体实施例并且不作为本申请的限制。在此使用时，单数形式的“一”、“一个”和“所述/该”也意图包括复数形式，除非上下文清楚指出另外的方式。还应明白术语“组成”和/或“包括”，当在该说明书中使用时，确定所述特征、整数、步骤、操作、元件和/或部件的存在，但不排除一个或更多其他的特征、整数、步骤、操作、元件、部件和/或组的存在或添加。在此使用时，术语“和/或”包括相关所列项目的任何及所有组合。
为了彻底理解本申请，将在下列的描述中提出详细的步骤以及详细的结构，以便阐释本申请提出的技术方案。本申请的较佳实施例详细描述如下，然而除了这些详细描述外，本申请还可以具有其他实施方式。
首先，参照图1来描述根据本申请实施例的样本分类方法。图1示出了根据本申请实施例的样本分类方法100的示意性流程图。如图1所示，样本分类方法100包括如下步骤：
在步骤S110，获取待分类样本的样本图像，在第一显示界面上显示所述样本图像。
在本申请的实施例中，在第一显示界面上显示待分类样本的样本图像。此处，术语“第一显示界面”是为了与后文将出现的用于显示待分类样本的备选分类结果的“第二显示界面”彼此区分而如此命名，没有其他限制意义。在本申请的实施例中，第一显示界面和第二显示界面可以是不同的显示界面，也可以是同一显示界面的不同状态。
在步骤S120，接收用于对所述样本图像对应的待分类样本进行分类的 分类指令。
在本申请的实施例中，可以接收用户通过输入装置(诸如显示界面的触摸屏或者鼠标键盘等)输入的对待分类样本进行分类的分类指令，诸如对待分类样本的样本图像执行拖拽指令、长按指令、点击指令等等。其中，拖拽指令可以包括：将待分类样本的样本图像拖拽预设距离，该预设距离可以为一个较小的数值，避免长距离拖拽产生的不便以及易出错问题。基于该分类指令，可以获取待分类样本的备选分类结果进行显示，如下面的步骤将描述的。
在步骤S130，基于所述分类指令在所述样本图像的周边区域显示所述待分类样本的备选分类结果，以供用户选择所述待分类样本的最终分类结果。
在本申请的实施例中，基于在步骤S120接收的分类指令，可以首先获取待分类样本的备选分类结果。在一个示例中，待分类样本的备选分类结果可以是基于对待分类样本的样本图像进行自动识别而得到的。在本申请的实施例中，在待分类样本的样本图像的周边区域显示待分类样本的备选分类结果，此处，样本图像的周边区域可以理解为：以所述样本图像为中心，在预定距离内围绕所述样本图像的区域，下面结合图2来示例性地描述。
图2示出了根据本申请实施例的样本分类方法针对待分类样本显示备选分类结果的一个示例。如图2所示，显示界面上包括多个待分类样本各自的样本图像，针对其中一个样本图像S1，在接收到分类指令后，在该样本图像S1的周边区域显示了其备选分类结果，包括：嗜碱性粒细胞、中性分叶核粒细胞、中性杆状核粒细胞、涂抹细胞以及沉渣。在图2所示的示例中，将样本图像S1对应的待分类样本的备选分类结果显示在以样本图像S1为中心、围绕样本图像S1的圆形区域内，应理解，这仅是示例性的，根据本申请的教导，将待分类样本的备选分类结果显示在待分类样本的样本图像的周边区域不必一定是圆形区域，只要是在样本图像的附近区域，使得用户可以便利地、快捷地选择即可。总体上，与将样本图像拖动较远距离来进行分类的方式相比较，本申请提供的在待分类样本的样本图像的周边区域显示待分类样本的备选分类结果的方式可以使得用户快速从备选 分类结果中选择待分类样本的最终分类结果，提高样本分类的效率，且用户操作简单，减少发生分类错误的可能性。
在本申请的实施例中，可以将待分类样本的备选分类结果显示在第二显示界面上，如前所述的，显示待分类样本的样本图像的第一显示界面与显示待分类样本的备选分类结果的第二显示界面可以是不同的显示界面，也可以是同一显示界面的不同状态。
在本申请的一个实施例中，在样本图像的周边区域显示的待分类样本的备选分类结果可以包括对样本图像对应的待分类样本进行分类得到的至少一个分类结果，各个备选分类结果可以基于各自的分类置信度而排序显示。其中，一个分类结果的分类置信度是指待分类样本的真实分类结果为该分类结果的概率。例如，对于一个白细胞的样本图像进行分类后，得到的备选分类结果中的一个分类结果为中性粒细胞，分类置信度为90％，则表示该白细胞为中性粒细胞的概率为90％。在该实施例中，将各备选分类结果按照各自的分类置信度排序显示，由于一个分类结果的分类置信度越高，表示该分类结果越有可能是真实的分类结果，因此基于分类置信度排序显示的方式可以使得用户更快速地找到待分类样本的最终分类结果，提高分类效率。
在一个示例中，分类置信度越高，对应的分类结果的显示位置越靠近该样本图像，这样可以进一步提高用户找到并选择待分类样本的最终分类结果的速度。在另一个示例中，在样本图像的周边区域显示的待分类样本的备选分类结果可以包括至少两级分类结果，其中前一级分类结果中的每一个的分类置信度高于后一级分类结果中的每一个的分类置信度，可以基于用户指令逐级显示每一级分类结果，以用于基于至少一级分类结果获得所述待分类样本的最终分类结果。下面结合图3A和图3B来描述该示例。
图3A和图3B示出了根据本申请实施例的样本分类方法针对待分类样本显示备选分类结果的另一个示例，其中，图3A示出待分类样本的第一级分类结果，图3B示出待分类样本的第二级分类结果。在该示例中，不是一次性显示待分类样本的所有备选分类结果，而是基于分级菜单显示，其中第一级分类结果显示分类置信度最高的前几个分类结果，如图3A所示的嗜酸性、嗜碱性、中性粒、淋巴、单核；如果用户认为第一级分类结果 中不存在待分类样本的真实分类结果或者不能确定第一级分类结构是否存在待分类样本的真实分类结果，则可以触发弹出第二级分类结果，第二级分类结果显示分类置信度次高的几个分类结果，如图3B所示的，原始、中性晚幼、异型淋巴、幼单核。应理解，这仅是示例性的，在实际应用中，所显示的备选分类结果可以包括不止两级分类结果。
在本申请的实施例中，用户在触发后一级分类结果的显示时，可以对前一级分类结果中的任一个执行悬浮指令、长按指令或者点击指令等等。或者，可以在前一级分类结果中提供“其他”选项，以由用户通过“其他”选项触发后一级分类结果，如图3A中的“其他”选项所示的。在本申请的实施例中，备选分类结果中还可以包括“分类不明”选项，如图3B所示的，该选项可以用于用户无法确定最终分类结果的场景。悬浮指令包括将指针悬浮在所选择的对象的预设区域以选中该对象的选择方式。
在本申请的实施例中，前一级分类结果相对于后一级分类结果可以更靠近样本图像，例如后一级分类结果可以显示在前一级分类结果的周边区域，例如如图3B所示的，第二级分类结果显示在第一级分类结果的周边区域，这样的显示方式层次分明且便于用户更灵活的选择，例如从第二级分类结果再回到第一级分类结果中进行选择。当然，这仅是示例性的，也可以在显示后一级分类结果时不显示前一级分类结果，同时提供一个可以返回至前一级分类结果的选项。
在一个示例中，前述的至少两级分类结果中的第一级分类结果可以在第一圆周包围的至少部分区域内显示，对于所述至少两级分类结果中的其余级分类结果，后一级分类结果围绕前一级分类结果在第二圆周包围的至少部分区域内显示，第一圆周与第二圆周同心且第二圆周的半径大于第一圆周的半径，例如如图3B所示的。在该示例中，以圆周显示方式逐级显示分类结果，圆周形状的面积利用率高、便于分级且视觉整齐，使得用户在选择待分类样本的最终分类结果时视觉上清晰明了，便于快捷且准确地选择。
总体上，前述的基于分类置信度分级显示分类结果的示例使得用户无需一次从众多选项中选择待分类样本的最终分类结果，而是可能在部分较少的备选分类结果中找到最终分类结果，进一步提高分类效率。
在本申请的进一步的实施例中，无论是否分级显示，在基于分类置信 度显示备选分类结果的实施例中，可以进一步通过如下方式显示：具有最高分类置信度的分类结果以区分于其他分类置信度的分类结果的显示方式显示，或者，不同分类置信度的分类结果被区分显示。示例性地，所述区分显示包括以下中的至少一项：以不同颜色的图标显示不同分类置信度的分类结果；以不同大小的图标显示不同分类置信度的分类结果；以相对于所述样本图像不同距离的图标显示不同分类置信度的分类结果；以不同属性的文字显示不同分类置信度的分类结果；在具有不同形状的区域显示不同分类置信度的分类结果；将各分类结果的分类置信度的数值显示在相应的分类结果上。在该实施例中，将具有最高分类置信度的分类结果以区分于其他分类置信度的分类结果的显示方式显示，或者将具有不同分类置信度的分类结果彼此区分显示，可以进一步强化用户的视觉效果，使得用户能够更明显地看到分类置信度高的分类结果，不仅能够进一步提高分类效率，还可以因为这样的区分显示而进一步减少分类错误的可能性。
在本申请的又一个实施例中，步骤S130所显示的待分类样本的备选分类结果可以包括至少两级分类结果，其中至少一级分类结果被分为多个组，所述多个组包含了该级分类中所有可能的分类结果，前一级分类结果中任一分组所包括的分类项目为后一级分类结果，对所述备选分类结果的显示包括：基于用户指令逐级显示每一级分类结果，以用于基于至少一级分类结果获得所述待分类样本的最终分类结果。在该实施例中，不是一次性显示待分类样本的所有备选分类结果，而是基于分级菜单显示，其中第一级分类结果显示对待分类样本的分组结果，第二级分类结果显示第一级分类结果中任一个分组的下级结果，即该分组进一步包括的分组或分类结果，以此类推。下面结合图4A和图4B来描述该示例。
图4A和图4B示出了根据本申请实施例的样本分类方法针对待分类样本显示备选分类结果的再一个示例，其中，图4A示出待分类样本的第一级分类结果，图4B示出待分类样本的第二级分类结果。在该示例中，第一级分类结果显示待分类样本可能属于的组的类别，如图4A所示的嗜酸性组、嗜碱性组、中性粒组、单核组、淋巴组；用户可以基于第一级分类结果选择待分类样本所属的组作为待分类样本的最终分类结果，也可以继续触发第二级分类结果，精确选择第一级分类结果中某一组别的下一级所属 类别，如图4B所示的原始、中性晚幼、异型淋巴、幼单核。应理解，这仅是示例性的，在实际应用中，所显示的备选分类结果可以包括不止两级分类结果。
与前述实施例类似的，在本实施例中，用户在触发后一级分类结果的显示时，可以对前一级分类结果中的任一个执行悬浮指令、长按指令或者点击指令等等。或者，可以在前一级分类结果中提供“其他”选项，以由用户通过“其他”选项触发后一级分类结果，如图4A中的“其他组”选项所示的。在本申请的实施例中，备选分类结果中还可以包括“分类不明”选项，如图4B所示的，该选项可以用于用户无法确定最终分类结果的场景。
在本申请的实施例中，前一级分类结果相对于后一级分类结果可以更靠近样本图像，后一级分类结果可以显示在前一级分类结果的周边区域，例如如图4B所示的，第二级分类结果显示在第一级分类结果的周边区域，这样的显示方式层次分明且便于用户更灵活的选择，例如从第二级分类结果再回到第一级分类结果中进行选择。当然，这仅是示例性的，也可以在显示后一级分类结果时不显示前一级分类结果，同时提供一个可以返回至前一级分类结果的选项。
在一个示例中，前述的至少两级分类结果中的第一级分类结果可以在第一圆周包围的至少部分区域内显示，对于所述至少两级分类结果中的其余级分类结果，后一级分类结果围绕前一级分类结果在第二圆周包围的至少部分区域内显示，第一圆周与第二圆周同心且第二圆周的半径大于第一圆周的半径，例如如图4B所示的。在该示例中，以圆周显示方式逐级显示分类结果，圆周形状的面积利用率高、便于分级且视觉整齐，使得用户在选择待分类样本的最终分类结果时视觉上清晰明了，便于快捷且准确地选择。
在本申请的进一步的实施例中，前述结合图4A和图4B描述的备选分类结果中的至少两级分类结果中每一级分类结果所包含的各分类项目还可以按照分类置信度的高低区分显示。在该实施例中，每一级分类结果所包含的各分类项目表示待分类样本的一个可能的分组结果或分类结果，且该结果还具有一个置信度，各分类项目基于置信度高低彼此区分显示可以进一步强化用户的视觉效果，提高用户选择分类结果的效率。
总体上，前述的基于分类置信度和/或基于分组的分级显示分类结果的示例能够提升用户信息筛选的效率，只有在前一级没有提供目标分类结果时，才展开下一级，或者前一级对后一级提供了引导信息，使得用户不需要同时筛选大量信息，从而能够进一步提高分类效率。
在本申请的其他实施例中，步骤S120所接收的分类指令还可以包括输入分类关键字的指令；在该实施例中，步骤S130显示的待分类样本的备选分类结果可以包括基于所述分类关键字的分类结果。在该实施例中，根据用户输入的分类关键字显示相应的备选分类结果，也可以减少显示的备选分类结果的数量，使得用户快速选择待分类样本的最终分类结果。
在本申请的进一步的实施例中，步骤S130显示的待分类样本的备选分类结果还可以包括基于历史数据得到的分类结果，所述历史数据包括先前样本的最终分类结果。在该实施例中，用户针对先前样本的最终分类结果的选择作为历史数据被用于当前样本的分类，因而可以结合用户的历史操作对当前样本进行分类，可以进一步提高所显示的备选分类结果为正确分类结果(真实分类结果)的可能性，从而进一步提高分类效率。示例性地，相对于前述的基于分类置信度排序显示的分类结果，此处基于历史数据得到的分类结果的显示位置可以更靠近待分类样本的样本图像；或者，相对于此处基于历史数据得到的分类结果，前述的基于分类置信度排序显示的分类结果的显示位置更靠近待分类样本的样本图像。
现在返回参考图1，描述根据本申请实施例的样本分类方法100的后续步骤。
在步骤S140，接收用户对所述备选分类结果的选择指令，并基于所述选择指令获取所述待分类样本的最终分类结果。
在本申请的实施例中，基于在步骤S130所显示的待分类样本的备选分类结果，用户可以通过选择指令从中选择待分类样本的最终分类结果。示例性地，该选择指令可以包括对任一备选分类结果的长按指令、点击指令等等。基于用户的选择指令，可以获取待分类样本的最终分类结果。在一个示例中，可以显示出待分类样本的最终分类结果。
在本申请的进一步的实施例中，方法100还可以包括(未示出)：在得到当前待分类样本的最终分类结果之后，对与当前待分类样本位于同一初 始分组的其余样本执行以下操作中的任一项：自动显示所述同一初始分组的其余样本的备选分类结果；基于当前待分类样本的最终分类结果生成所述同一初始分组的其余样本的最终分类结果；其中，所述初始分组是基于对样本的初始分类结果而分组生成的，当前待分类样本为待进行重分类的样本，所述重分类是对所述初始分类结果的验证或修正，所述样本经重分类后得到的重分类结果为所述最终分类结果。
在该实施例中，本申请的样本分类方法100是用于对样本的重分类，也就是说，样本在重分类之前已经经过初始分类，并且多个样本可能被初始分类为同一组，在该情况下，对于同一组中任一样本按照根据本申请实施例的样本分类方法100重分类后，可以无需对该同一组中其他样本的样本图像输入分类指令就自动启动对其他样本的备选分类结果的呈现，也可以甚至无需再对其他样本重复根据本申请实施例的样本分类方法100的过程而直接将该组中已被重分类的一个样本的最终分类结果作为其他样本的最终分类结果，这可以进一步提高多个样本分类的效率。
基于上面的描述，根据本申请实施例的样本分类方法将待分类样本的备选分类结果显示在待分类样本的样本图像的周边区域，使得用户可以快速从备选分类结果中选择待分类样本的最终分类结果，提高样本分类的效率，且用户操作简单，减少发生分类错误的可能性。此外，根据本申请实施例的样本分类方法还可以将待分类样本的备选分类结果进行分级显示，能够提升用户信息筛选的效率，只有在前一级没有提供目标分类结果时，才展开下一级，或者前一级对后一级提供了引导信息，使得用户不需要同时筛选大量信息，从而能够进一步提高分类效率。此外，根据本申请实施例的样本分类方法还可以将待分类样本的具有不同分类置信度的备选分类结果进行区分显示，可以进一步强化用户的视觉效果，使得用户能够更明显地看到分类置信度高的分类结果，不仅能够进一步提高分类效率，还可以因为这样的区分显示而进一步减少分类错误的可能性。
以上示例性地示出了根据本申请实施例的样本分类方法。下面结合图5描述根据申请另一方面提供的样本分类装置。图5示出了根据本申请实施例的样本分类装置500的示意性框图。如图5所示，样本分类装置500可以包括存储器510、处理器520和显示器530。其中，存储器510存储用 于实现根据本申请实施例的样本分类方法100中的相应步骤的程序。处理器520用于运行存储器510中存储的程序，以执行根据本申请实施例的样本分类方法100的相应步骤，显示器530用于基于处理器520的控制显示信息。本领域技术人员可以结合前文的描述理解该样本分类装置500的结构及操作，为了简洁，此处仅描述处理器520的主要操作，对于上文中的一些细节不再赘述。
在本申请的一个实施例中，所述计算机程序在被处理器520运行时，使得处理器520执行以下步骤：获取所述待分类样本的样本图像，并由显示器530在第一显示界面上显示所述样本图像；接收用于对所述样本图像对应的待分类样本进行分类的分类指令；基于所述分类指令控制在所述样本图像的周边区域显示所述待分类样本的备选分类结果，以供用户选择所述待分类样本的最终分类结果；接收用户对所述备选分类结果的选择指令，并基于所述选择指令获取所述待分类样本的最终分类结果。
在本申请的一个实施例中，所述备选分类结果是通过处理器520对所述样本图像进行识别而得到的。
在本申请的一个实施例中，所述备选分类结果包括对所述样本图像对应的待分类样本进行分类得到的至少一个分类结果，各个所述备选分类结果基于各自的分类置信度而排序显示。
在本申请的一个实施例中，所述排序显示包括：分类置信度越高，对应的分类结果的显示位置越靠近所述样本图像。
在本申请的一个实施例中，所述备选分类结果包括至少两级分类结果，其中前一级分类结果中的每一个的分类置信度高于后一级分类结果中的每一个的分类置信度，所述排序显示包括：基于用户指令逐级显示每一级分类结果，以用于基于至少一级分类结果获得所述待分类样本的最终分类结果。
在本申请的一个实施例中，在显示器530上显示所述备选分类结果或显示所述备选分类结果中的任一级分类结果时，具有最高分类置信度的分类结果以区分于其他分类置信度的分类结果的显示方式显示，或者，不同分类置信度的分类结果被区分显示。
在本申请的一个实施例中，所述区分显示包括以下中的至少一项：以 不同颜色的图标显示不同分类置信度的分类结果；以不同大小的图标显示不同分类置信度的分类结果；以相对于所述样本图像不同距离的图标显示不同分类置信度的分类结果；以不同属性的文字显示不同分类置信度的分类结果；在具有不同形状的区域显示不同分类置信度的分类结果；将各分类结果的分类置信度的数值显示在相应的分类结果上。
在本申请的一个实施例中，所述备选分类结果包括至少两级分类结果，其中至少一级分类结果被分为多个组，所述多个组包含了该级分类中所有可能的分类结果，前一级分类结果中任一分组所包括的分类项目为后一级分类结果，显示器530对所述备选分类结果的显示包括：基于用户指令逐级显示每一级分类结果，以用于基于至少一级分类结果获得所述待分类样本的最终分类结果。
在本申请的一个实施例中，所述至少两级分类结果中每一级分类结果所包含的各分类项目按照分类置信度的高低区分显示。
在本申请的一个实施例中，所述后一级分类结果显示在所述前一级分类结果的周边区域。
在本申请的一个实施例中，所述至少两级分类结果中的第一级分类结果在第一圆周包围的至少部分区域内显示，对于所述至少两级分类结果中的其余级分类结果，所述后一级分类结果围绕所述前一级分类结果在第二圆周包围的至少部分区域内显示，所述第一圆周与所述第二圆周同心且所述第二圆周的半径大于所述第一圆周的半径。
在本申请的一个实施例中，所述前一级分类结果相对于所述后一级分类结果更靠近所述样本图像。
在本申请的一个实施例中，所述备选分类结果显示在第二显示界面上，所述第一显示界面和所述第二显示界面为不同的显示界面。
在本申请的一个实施例中，所述备选分类结果和所述样本图像在所述第一显示界面的不同状态下显示。
在本申请的一个实施例中，所述分类指令包括对所述样本图像执行的以下任一项指令：拖拽指令、长按指令以及点击指令。
在本申请的一个实施例中，所述分类指令包括输入分类关键字的指令，所述备选分类结果包括基于所述分类关键字的分类结果。
在本申请的一个实施例中，所述用户指令包括所述分类指令以及对所述前一级分类结果中的任一个执行的以下任一项指令：悬浮指令、长按指令以及点击指令。
在本申请的一个实施例中，显示器530还用于：显示所述待分类样本的最终分类结果。
在本申请的一个实施例中，针对所述待分类样本显示的备选分类结果还包括基于历史数据得到的分类结果，所述历史数据包括先前样本的最终分类结果。
在本申请的一个实施例中，相对于基于分类置信度排序显示的分类结果，所述基于历史数据得到的分类结果的显示位置更靠近所述样本图像；或者相对于所述基于历史数据得到的分类结果，基于分类置信度排序显示的分类结果的显示位置更靠近所述样本图像。
在本申请的一个实施例中，所述计算机程序在被处理器520运行时，还使得处理器520执行以下步骤：在得到所述待分类样本的最终分类结果之后，对与所述待分类样本位于同一初始分组的其余样本执行以下操作中的任一项：自动在显示器530上显示所述同一初始分组的其余样本的备选分类结果；基于所述待分类样本的最终分类结果生成所述同一初始分组的其余样本的最终分类结果；其中，所述初始分组是基于对样本的初始分类结果而分组生成的，所述待分类样本为待进行重分类的样本，所述重分类是对所述初始分类结果的验证或修正，所述样本经重分类后得到的重分类结果为所述最终分类结果。
在本申请的一个实施例中，所述周边区域包括以所述样本图像为中心，在预定距离内围绕所述样本图像的区域。
在本申请的一个实施例中，所述备选分类结果包括至少两级分类结果，所述至少两级分类结果基于用户指令逐级显示，所述待分类样本的最终分类结果基于至少一级分类结果而得到。
此外，根据本申请实施例，还提供了一种存储介质，在存储介质上存储了程序指令，在程序指令被计算机或处理器运行时用于执行本申请实施例的样本分类方法的相应步骤。存储介质例如可以包括智能电话的存储卡、平板电脑的存储部件、个人计算机的硬盘、只读存储器(ROM)、可擦除可 编程只读存储器(EPROM)、便携式紧致盘只读存储器(CD-ROM)、USB存储器、或者上述存储介质的任意组合。计算机可读存储介质可以是一个或多个计算机可读存储介质的任意组合。
此外，根据本申请实施例，还提供了一种计算机程序，该计算机程序可以存储在云端或本地的存储介质上。在该计算机程序被计算机或处理器运行时用于执行本申请实施例的样本分类方法的相应步骤。
下面参考图6描述根据本申请实施例的样本分类方法的应用场景的示例。图6示出了阅片机的结构示意图。如图6所示，阅片机是一种医疗设备，阅片机包括相对设置的物镜601和阅片平台602，其中，阅片平台602上放置有玻片603，该玻片603上滴有样本，例如经染色后的血液样本或体液样本等，该物镜601与相机连接，从而使得相机可以通过该物镜601拍摄到玻片603上的样本，得到样本图像。
针对该样本图像，阅片机会通过对样本图像的自动识别而在人机交互装置(未示出)上输出与该样本图像对应的样本的初始分类结果，根据本申请实施例的样本分类方法可用于对经初始分类的样本进行重分类，用户通过操作该人机交互装置可调用本申请实施例所提供的样本分类方法，并从在样本图像的周边区域显示的样本的备选分类结果中快速高效地选择样本的最终分类结果，如前文所描述的。
基于上面的描述，根据本申请实施例的样本分类方法、装置和存储介质将待分类样本的备选分类结果显示在待分类样本的样本图像的周边区域，使得用户可以快速从备选分类结果中选择待分类样本的最终分类结果，提高样本分类的效率，且用户操作简单，减少发生分类错误的可能性。此外，根据本申请实施例的样本分类方法、装置和存储介质还可以将待分类样本的备选分类结果进行分级显示，能够提升用户信息筛选的效率，只有在前一级没有提供目标分类结果时，才展开下一级，或者前一级对后一级提供了引导信息，使得用户不需要同时筛选大量信息，从而能够进一步提高分类效率。此外，根据本申请实施例的样本分类方法、装置和存储介质还可以将待分类样本的具有不同分类置信度的备选分类结果进行区分显示，可以进一步强化用户的视觉效果，使得用户能够更明显地看到分类置信度高的分类结果，不仅能够进一步提高分类效率，还可以因为这样的区分显示而进一步 减少分类错误的可能性。
尽管这里已经参考附图描述了示例实施例，应理解上述示例实施例仅仅是示例性的，并且不意图将本申请的范围限制于此。本领域普通技术人员可以在其中进行各种改变和修改，而不偏离本申请的范围和精神。所有这些改变和修改意在被包括在所附权利要求所要求的本申请的范围之内。
本领域普通技术人员可以意识到，结合本文中所公开的实施例描述的各示例的单元及算法步骤，能够以电子硬件、或者计算机软件和电子硬件的结合来实现。这些功能究竟以硬件还是软件方式来执行，取决于技术方案的特定应用和设计约束条件。专业技术人员可以对每个特定的应用来使用不同方法来实现所描述的功能，但是这种实现不应认为超出本申请的范围。
在本申请所提供的几个实施例中，应该理解到，所揭露的装置和方法，可以通过其他的方式实现。例如，以上所描述的装置实施例仅仅是示意性的，例如，单元的划分，仅仅为一种逻辑功能划分，实际实现时可以有另外的划分方式，例如多个单元或组件可以结合或者可以集成到另一个装置，或一些特征可以忽略，或不执行。
在此处所提供的说明书中，说明了大量具体细节。然而，能够理解，本申请的实施例可以在没有这些具体细节的情况下实践。在一些实例中，并未详细示出公知的方法、结构和技术，以便不模糊对本说明书的理解。
类似地，应当理解，为了精简本申请并帮助理解各个发明方面中的一个或多个，在对本申请的示例性实施例的描述中，本申请的各个特征有时被一起分组到单个实施例、图、或者对其的描述中。然而，并不应将该本申请的方法解释成反映如下意图：即所要求保护的本申请要求比在每个权利要求中所明确记载的特征更多的特征。更确切地说，如相应的权利要求书所反映的那样，其发明点在于可以用少于某个公开的单个实施例的所有特征的特征来解决相应的技术问题。因此，遵循具体实施方式的权利要求书由此明确地并入该具体实施方式，其中每个权利要求本身都作为本申请的单独实施例。
本领域的技术人员可以理解，除了特征之间相互排斥之外，可以采用任何组合对本说明书(包括伴随的权利要求、摘要和附图)中公开的所有 特征以及如此公开的任何方法或者装置的所有过程或单元进行组合。除非另外明确陈述，本说明书(包括伴随的权利要求、摘要和附图)中公开的每个特征可以由提供相同、等同或相似目的的替代特征来代替。
此外，本领域的技术人员能够理解，尽管在此的一些实施例包括其他实施例中所包括的某些特征而不是其他特征，但是不同实施例的特征的组合意味着处于本申请的范围之内并且形成不同的实施例。例如，在权利要求书中，所要求保护的实施例的任意之一都可以以任意的组合方式来使用。
本申请的各个部件实施例可以以硬件实现，或者以在一个或者多个处理器上运行的软件模块实现，或者以它们的组合实现。本领域的技术人员应当理解，可以在实践中使用微处理器或者数字信号处理器(DSP)来实现根据本申请实施例的一些模块的一些或者全部功能。本申请还可以实现为用于执行这里所描述的方法的一部分或者全部的装置程序(例如，计算机程序和计算机程序产品)。这样的实现本申请的程序可以存储在计算机可读介质上，或者可以具有一个或者多个信号的形式。这样的信号可以从因特网网站上下载得到，或者在载体信号上提供，或者以任何其他形式提供。
应该注意的是上述实施例对本申请进行说明而不是对本申请进行限制，并且本领域技术人员在不脱离所附权利要求的范围的情况下可设计出替换实施例。在权利要求中，不应将位于括号之间的任何参考符号构造成对权利要求的限制。本申请可以借助于包括有若干不同元件的硬件以及借助于适当编程的计算机来实现。在列举了若干装置的单元权利要求中，这些装置中的若干个可以是通过同一个硬件项来具体体现。单词第一、第二、以及第三等的使用不表示任何顺序。可将这些单词解释为名称。
以上，仅为本申请的具体实施方式或对具体实施方式的说明，本申请的保护范围并不局限于此，任何熟悉本技术领域的技术人员在本申请揭露的技术范围内，可轻易想到变化或替换，都应涵盖在本申请的保护范围之内。本申请的保护范围应以权利要求的保护范围为准。

Claims (47)

  1. 一种样本分类方法，其特征在于，所述方法包括：
    获取待分类样本的样本图像，在第一显示界面上显示所述样本图像；
    接收用于对所述样本图像对应的待分类样本进行分类的分类指令；
    基于所述分类指令在所述样本图像的周边区域显示所述待分类样本的备选分类结果，以供用户选择所述待分类样本的最终分类结果；
    接收用户对所述备选分类结果的选择指令，并基于所述选择指令获取所述待分类样本的最终分类结果。
  2. 根据权利要求1所述的方法，其特征在于，所述备选分类结果是基于对所述样本图像的自动识别而得到的。
  3. 根据权利要求1所述的方法，其特征在于，所述备选分类结果包括对所述样本图像对应的待分类样本进行分类得到的至少一个分类结果，各个所述备选分类结果基于各自的分类置信度而排序显示。
  4. 根据权利要求3所述的方法，其特征在于，所述排序显示包括：分类置信度越高，对应的分类结果的显示位置越靠近所述样本图像。
  5. 根据权利要求3所述的方法，其特征在于，所述备选分类结果包括至少两级分类结果，其中前一级分类结果中的每一个的分类置信度高于后一级分类结果中的每一个的分类置信度，所述排序显示包括：
    基于用户指令逐级显示每一级分类结果，以用于基于至少一级分类结果获得所述待分类样本的最终分类结果。
  6. 根据权利要求3-5中的任一项所述的方法，其特征在于，在显示所述备选分类结果或显示所述备选分类结果中的任一级分类结果时，具有最高分类置信度的分类结果以区分于其他分类置信度的分类结果的显示方式显示，或者，不同分类置信度的分类结果被区分显示。
  7. 根据权利要求6所述的方法，其特征在于，所述区分显示包括以下中的至少一项：
    以不同颜色的图标显示不同分类置信度的分类结果；
    以不同大小的图标显示不同分类置信度的分类结果；
    以相对于所述样本图像不同距离的图标显示不同分类置信度的分类结果；
    以不同属性的文字显示不同分类置信度的分类结果；
    在具有不同形状的区域显示不同分类置信度的分类结果；
    将各分类结果的分类置信度的数值显示在相应的分类结果上。
  8. 根据权利要求1所述的方法，其特征在于，所述备选分类结果包括至少两级分类结果，其中至少一级分类结果被分为多个组，所述多个组包含了该级分类中所有可能的分类结果，前一级分类结果中任一分组所包括的分类项目为后一级分类结果，对所述备选分类结果的显示包括：
    基于用户指令逐级显示每一级分类结果，以用于基于至少一级分类结果获得所述待分类样本的最终分类结果。
  9. 根据权利要求8所述的方法，其特征在于，所述至少两级分类结果中每一级分类结果所包含的各分类项目按照分类置信度的高低区分显示。
  10. 根据权利要求5或8所述的方法，其特征在于，所述后一级分类结果显示在所述前一级分类结果的周边区域。
  11. 根据权利要求10所述的方法，其特征在于，所述至少两级分类结果中的第一级分类结果在第一圆周包围的至少部分区域内显示，对于所述至少两级分类结果中的其余级分类结果，所述后一级分类结果围绕所述前一级分类结果在第二圆周包围的至少部分区域内显示，所述第一圆周与所述第二圆周同心且所述第二圆周的半径大于所述第一圆周的半径。
  12. 根据权利要求5或8所述的方法，其特征在于，所述前一级分类结果相对于所述后一级分类结果更靠近所述样本图像。
  13. 根据权利要求1所述的方法，其特征在于，所述备选分类结果显示在第二显示界面上，所述第一显示界面和所述第二显示界面为不同的显示界面。
  14. 根据权利要求1所述的方法，其特征在于，所述备选分类结果和所述样本图像在所述第一显示界面的不同状态下显示。
  15. 根据权利要求1所述的方法，其特征在于，所述分类指令包括对所述样本图像执行的以下任一项指令：拖拽指令、长按指令以及点击指令。
  16. 根据权利要求1或2所述的方法，其特征在于，所述分类指令包括输入分类关键字的指令，所述备选分类结果包括基于所述分类关键字的分类结果。
  17. 根据权利要求5或8所述的方法，其特征在于，所述用户指令包括所述分类指令以及对所述前一级分类结果中的任一个执行的以下任一项指令：悬浮指令、长按指令以及点击指令。
  18. 根据权利要求1所述的方法，其特征在于，所述方法还包括：
    显示所述待分类样本的最终分类结果。
  19. 根据权利要求3所述的方法，其特征在于，针对所述待分类样本显示的备选分类结果还包括基于历史数据得到的分类结果，所述历史数据包括先前样本的最终分类结果。
  20. 根据权利要求19所述的方法，其特征在于，相对于基于分类置信度排序显示的分类结果，所述基于历史数据得到的分类结果的显示位置更靠近所述样本图像；或者，相对于所述基于历史数据得到的分类结果，基于分类置信度排序显示的分类结果的显示位置更靠近所述样本图像。
  21. 根据权利要求18或19所述的方法，其特征在于，所述方法还包括：
    在得到所述待分类样本的最终分类结果之后，对与所述待分类样本位于同一初始分组的其余样本执行以下操作中的任一项：
    自动显示所述同一初始分组的其余样本的备选分类结果；
    基于所述待分类样本的最终分类结果生成所述同一初始分组的其余样本的最终分类结果；
    其中，所述初始分组是基于对样本的初始分类结果而分组生成的，所述待分类样本为待进行重分类的样本，所述重分类是对所述初始分类结果的验证或修正，所述样本经重分类后得到的重分类结果为所述最终分类结果。
  22. 根据权利要求1所述的方法，其特征在于，所述周边区域包括以所述样本图像为中心，在预定距离内围绕所述样本图像的区域。
  23. 根据权利要求1所述的方法，其特征在于，所述备选分类结果包括至少两级分类结果，所述至少两级分类结果基于用户指令逐级显示，所述待分类样本的最终分类结果基于至少一级分类结果而得到。
  24. 一种样本分类装置，其特征在于，所述装置包括存储器、处理器和显示器，其中，所述存储器用于存储程序和待分类样本的样本图像，所 述显示器用于基于所述处理器的控制进行显示，所述处理器用于运行所述存储器中的程序以执行以下步骤：
    获取所述待分类样本的样本图像，并由所述显示器在第一显示界面上显示所述样本图像；
    接收用于对所述样本图像对应的待分类样本进行分类的分类指令；
    基于所述分类指令控制在所述样本图像的周边区域显示所述待分类样本的备选分类结果，以供用户选择所述待分类样本的最终分类结果；
    接收用户对所述备选分类结果的选择指令，并基于所述选择指令获取所述待分类样本的最终分类结果。
  25. 根据权利要求24所述的装置，其特征在于，所述备选分类结果是通过所述处理器对所述样本图像进行识别而得到的。
  26. 根据权利要求24所述的装置，其特征在于，所述备选分类结果包括对所述样本图像对应的待分类样本进行分类得到的至少一个分类结果，各个所述备选分类结果基于各自的分类置信度而排序显示。
  27. 根据权利要求26所述的装置，其特征在于，所述排序显示包括：分类置信度越高，对应的分类结果的显示位置越靠近所述样本图像。
  28. 根据权利要求26所述的装置，其特征在于，所述备选分类结果包括至少两级分类结果，其中前一级分类结果中的每一个的分类置信度高于后一级分类结果中的每一个的分类置信度，所述排序显示包括：
    基于用户指令逐级显示每一级分类结果，以用于基于至少一级分类结果获得所述待分类样本的最终分类结果。
  29. 根据权利要求26-28中的任一项所述的装置，其特征在于，在所述显示器上显示所述备选分类结果或显示所述备选分类结果中的任一级分类结果时，具有最高分类置信度的分类结果以区分于其他分类置信度的分类结果的显示方式显示，或者，不同分类置信度的分类结果被区分显示。
  30. 根据权利要求29所述的装置，其特征在于，所述区分显示包括以下中的至少一项：
    以不同颜色的图标显示不同分类置信度的分类结果；
    以不同大小的图标显示不同分类置信度的分类结果；
    以相对于所述样本图像不同距离的图标显示不同分类置信度的分类 结果；
    以不同属性的文字显示不同分类置信度的分类结果；
    在具有不同形状的区域显示不同分类置信度的分类结果；
    将各分类结果的分类置信度的数值显示在相应的分类结果上。
  31. 根据权利要求24所述的装置，其特征在于，所述备选分类结果包括至少两级分类结果，其中至少一级分类结果被分为多个组，所述多个组包含了该级分类中所有可能的分类结果，前一级分类结果中任一分组所包括的分类项目为后一级分类结果，所述显示器对所述备选分类结果的显示包括：
    基于用户指令逐级显示每一级分类结果，以用于基于至少一级分类结果获得所述待分类样本的最终分类结果。
  32. 根据权利要求31所述的装置，其特征在于，所述至少两级分类结果中每一级分类结果所包含的各分类项目按照分类置信度的高低区分显示。
  33. 根据权利要求28或31所述的装置，其特征在于，所述后一级分类结果显示在所述前一级分类结果的周边区域。
  34. 根据权利要求33所述的装置，其特征在于，所述至少两级分类结果中的第一级分类结果在第一圆周包围的至少部分区域内显示，对于所述至少两级分类结果中的其余级分类结果，所述后一级分类结果围绕所述前一级分类结果在第二圆周包围的至少部分区域内显示，所述第一圆周与所述第二圆周同心且所述第二圆周的半径大于所述第一圆周的半径。
  35. 根据权利要求28或31所述的装置，其特征在于，所述前一级分类结果相对于所述后一级分类结果更靠近所述样本图像。
  36. 根据权利要求24所述的装置，其特征在于，所述备选分类结果显示在第二显示界面上，所述第一显示界面和所述第二显示界面为不同的显示界面。
  37. 根据权利要求24所述的装置，其特征在于，所述备选分类结果和所述样本图像在所述第一显示界面的不同状态下显示。
  38. 根据权利要求24所述的装置，其特征在于，所述分类指令包括对所述样本图像执行的以下任一项指令：拖拽指令、长按指令以及点击指令。
  39. 根据权利要求24或25所述的装置，其特征在于，所述分类指令 包括输入分类关键字的指令，所述备选分类结果包括基于所述分类关键字的分类结果。
  40. 根据权利要求28或31所述的装置，其特征在于，所述用户指令包括所述分类指令以及对所述前一级分类结果中的任一个执行的以下任一项指令：悬浮指令、长按指令以及点击指令。
  41. 根据权利要求24所述的装置，其特征在于，所述显示器还用于：
    显示所述待分类样本的最终分类结果。
  42. 根据权利要求26所述的装置，其特征在于，针对所述待分类样本显示的备选分类结果还包括基于历史数据得到的分类结果，所述历史数据包括先前样本的最终分类结果。
  43. 根据权利要求42所述的装置，其特征在于，相对于基于分类置信度排序显示的分类结果，所述基于历史数据得到的分类结果的显示位置更靠近所述样本图像；或者相对于所述基于历史数据得到的分类结果，基于分类置信度排序显示的分类结果的显示位置更靠近所述样本图像。
  44. 根据权利要求41或42所述的装置，其特征在于，所述处理器还用于：
    在得到所述待分类样本的最终分类结果之后，对与所述待分类样本位于同一初始分组的其余样本执行以下操作中的任一项：
    自动在所述显示器上显示所述同一初始分组的其余样本的备选分类结果；
    基于所述待分类样本的最终分类结果生成所述同一初始分组的其余样本的最终分类结果；
    其中，所述初始分组是基于对样本的初始分类结果而分组生成的，所述待分类样本为待进行重分类的样本，所述重分类是对所述初始分类结果的验证或修正，所述样本经重分类后得到的重分类结果为所述最终分类结果。
  45. 根据权利要求24所述的装置，其特征在于，所述周边区域包括以所述样本图像为中心，在预定距离内围绕所述样本图像的区域。
  46. 根据权利要求24所述的装置，其特征在于，所述备选分类结果包括至少两级分类结果，所述至少两级分类结果基于用户指令逐级显示，所 述待分类样本的最终分类结果基于至少一级分类结果而得到。
  47. 一种存储介质，其特征在于，所述存储介质上存储有计算机程序，所述计算机程序在运行时执行如权利要求1-23中的任一项所述的样本分类方法。
PCT/CN2020/105878 2020-07-30 2020-07-30 样本分类方法、装置和存储介质 WO2022021224A1 (zh)

Priority Applications (2)

Application Number Priority Date Filing Date Title
CN202080102603.6A CN115867945A (zh) 2020-07-30 2020-07-30 样本分类方法、装置和存储介质
PCT/CN2020/105878 WO2022021224A1 (zh) 2020-07-30 2020-07-30 样本分类方法、装置和存储介质

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
PCT/CN2020/105878 WO2022021224A1 (zh) 2020-07-30 2020-07-30 样本分类方法、装置和存储介质

Publications (1)

Publication Number Publication Date
WO2022021224A1 true WO2022021224A1 (zh) 2022-02-03

Family

ID=80037383

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2020/105878 WO2022021224A1 (zh) 2020-07-30 2020-07-30 样本分类方法、装置和存储介质

Country Status (2)

Country Link
CN (1) CN115867945A (zh)
WO (1) WO2022021224A1 (zh)

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20140341476A1 (en) * 2013-05-15 2014-11-20 Google Inc. Associating classifications with images
CN105308628A (zh) * 2013-05-08 2016-02-03 细胞视觉公司 用于红血细胞分析的图形用户界面
WO2017055412A1 (en) * 2015-09-30 2017-04-06 Siemens Healthcare Gmbh Method and system for classification of endoscopic images using deep decision networks
CN109754398A (zh) * 2019-01-14 2019-05-14 武汉兰丁医学高科技有限公司 一种基于计算机的骨髓白细胞自动标记方法及***
CN109934194A (zh) * 2019-03-20 2019-06-25 深圳市网心科技有限公司 图片分类方法、边缘设备、***及存储介质
CN110428007A (zh) * 2019-08-01 2019-11-08 科大讯飞(苏州)科技有限公司 X光图像目标检测方法、装置及设备

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105308628A (zh) * 2013-05-08 2016-02-03 细胞视觉公司 用于红血细胞分析的图形用户界面
US20140341476A1 (en) * 2013-05-15 2014-11-20 Google Inc. Associating classifications with images
WO2017055412A1 (en) * 2015-09-30 2017-04-06 Siemens Healthcare Gmbh Method and system for classification of endoscopic images using deep decision networks
CN109754398A (zh) * 2019-01-14 2019-05-14 武汉兰丁医学高科技有限公司 一种基于计算机的骨髓白细胞自动标记方法及***
CN109934194A (zh) * 2019-03-20 2019-06-25 深圳市网心科技有限公司 图片分类方法、边缘设备、***及存储介质
CN110428007A (zh) * 2019-08-01 2019-11-08 科大讯飞(苏州)科技有限公司 X光图像目标检测方法、装置及设备

Also Published As

Publication number Publication date
CN115867945A (zh) 2023-03-28

Similar Documents

Publication Publication Date Title
US10460256B2 (en) Interactive performance visualization of multi-class classifier
US11361526B2 (en) Content-aware selection
US9886669B2 (en) Interactive visualization of machine-learning performance
US9036888B2 (en) Systems and methods for performing quality review scoring of biomarkers and image analysis methods for biological tissue
US20130290225A1 (en) Systems and methods for selecting and analyzing particles in a biological tissue
US8737709B2 (en) Systems and methods for performing correlation analysis on clinical outcome and characteristics of biological tissue
AU2013255961B2 (en) Systems and methods for performing quality review scoring of biomarkers and image analysis methods for biological tissue
CN103518183A (zh) 图形对象分类
US11972097B2 (en) Selecting content in ink documents using a hierarchical data structure
WO2021231040A1 (en) Descriptive insight generation and presentation system
JP2021002354A (ja) 表示制御装置、表示制御方法及び表示制御プログラム
WO2020242341A1 (ru) Метод для выделения и классификации типов клеток крови с помощью глубоких сверточных нейронных сетей
WO2022021224A1 (zh) 样本分类方法、装置和存储介质
US20180052589A1 (en) User interface with tag in focus
JP2023070496A (ja) データ処理システム
CN114298958A (zh) 超声成像***和超声图像分析方法
US20230127415A1 (en) Information processing device, information processing method, and program
US20130286043A1 (en) Systems and methods for analyzing biomarker co-localization in a biological tissue
EP4123501A1 (en) Medical diagnosis assistance system and method
US11867929B2 (en) Information processing apparatus and non-transitory computer readable medium
KR20230056175A (ko) 의료 영상 분석 방법
CN115270745A (zh) 智能在线绘图方法及***
Lindlbauer Perceptual Grouping of Digital Sketches

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 20947351

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 20947351

Country of ref document: EP

Kind code of ref document: A1