Method, apparatus and device for processing empirical screenshots for medical rating

Abstract

The present invention relates to the technical field of image processing, and discloses a processing method, device and equipment for empirical screenshots of medical rating indicators. The present invention performs similarity matching between each target indicator empirical screenshot and the medical software function interface screenshot carrying picture feature information in the picture feature information library to obtain target picture feature information; calculates the picture scaling ratio according to the picture resolution of the target indicator empirical screenshot and the corresponding target picture feature information to calculate the target picture annotation information, determines and identifies the picture annotation area for the target picture annotation information to obtain the corresponding target annotation content text; furthermore, screens out the scene text from each target annotation content text, and performs scene matching on each target indicator empirical screenshot according to the scene text to realize the judgment of whether the target indicator empirical screenshots belong to the same scene. The present invention can effectively improve the recognition accuracy and processing efficiency.

Description

Technical Field

[0001] The present invention relates to the technical field of image processing, and in particular, to a method, device and equipment for processing empirical screenshots of medical rating indicators. Background Art

[0002] At present, when medical institutions conduct the grading evaluation of the application level of electronic medical record devices, they need to submit empirical materials on the basic items of device functions and empirical materials on the selected items of device functions. These two empirical materials are PDF documents, and the document content mainly includes index requirements, specific implementation methods of the index, and empirical screenshots of the index. Among them, the empirical screenshot of the index refers to a screenshot of the function interface related to the medical software involved in implementing the index. There are strict requirements for submitting empirical screenshots of the index, specifically: the function interface in the empirical screenshot of the index must be the recent software function interface; all empirical screenshots of the same index must belong to the same scenario (here refers to satisfying the same patient and the same operation time period); the privacy information of patients cannot appear in the empirical screenshots of the index.

[0003] According to the requirements of the grading evaluation of the application level of electronic medical record devices, the two empirical materials usually contain more than 1000 empirical screenshots of the index. If the empirical screenshots of each index are processed manually to meet the corresponding submission requirements, it will take a lot of time and it is easy to cancel the rating qualification due to processing errors.

[0004] Currently, when using the existing image intelligent matching technology to process the empirical screenshots of the index, it is usually to recognize the text content in the empirical screenshots of the index, and then compare the recognized result with the text content to be matched, so as to determine whether multiple empirical screenshots of the same index belong to the same scenario. However, the process of image matching based on text content is time-consuming, and when there are a large number of elements such as tables in the screenshots of the function interface related to the medical software, it is easy to have recognition errors and low efficiency. Summary of the Invention

[0005] The present invention provides a method, device and equipment for processing empirical screenshots of medical rating indicators, and solves the technical problems of high recognition error rate and low efficiency when the existing image intelligent matching technology processes the empirical screenshots of the index.

[0006] The first aspect of the present invention provides a method for processing empirical screenshots of medical rating indicators, including:

[0007] Taking the screenshot of the function interface of the target medical software as the target empirical screenshot of the index, and constructing the corresponding target empirical screenshot set of the index;

[0008] Performing similarity matching on each target indicator empirical screenshot in the target indicator empirical screenshot set and a medical software function interface screenshot carrying image feature information in the image feature information library, and obtaining corresponding image feature information as corresponding target image feature information according to the matching result; the image feature information includes image annotation information and image resolution;

[0009] Obtaining a picture resolution of the target indicator demonstration screenshot as the corresponding target picture resolution, and calculating a picture scaling ratio of the corresponding target indicator demonstration screenshot according to the target picture resolution and the corresponding target picture feature information;

[0010] Calculate the target image annotation information of the corresponding target indicator empirical screenshot according to the calculated image zoom ratio and the image annotation information in the corresponding target image feature information, identify the annotation coordinate area of the target image annotation information, and obtain the corresponding target annotation content text according to the annotation coordinate area;

[0011] The target image annotation information and the target annotation content text are used as the feature information to be matched of the target indicator empirical screenshot, and the scene text is filtered out from each target annotation content text according to the feature information to be matched of each target indicator empirical screenshot;

[0012] Scene matching is performed on each target indicator verification screenshot in the target indicator verification screenshot set according to the scene text. If the target image annotation information of the target indicator verification screenshot meets the preset same scene condition and the corresponding target annotation content text matches the scene text, it is determined that the corresponding target indicator verification screenshots belong to the same scene.

[0013] According to an achievable manner of the first aspect of the present invention, the picture annotation information includes annotation coordinates, and the target picture annotation information of the corresponding target indicator empirical screenshot is calculated according to the calculated picture scaling ratio and the picture annotation information in the corresponding target picture feature information, including:

[0014] The product of the calculated image scaling ratio and the corresponding annotation coordinates is used as the target annotation coordinates of the corresponding target indicator empirical screenshot.

[0015] According to an achievable manner of the first aspect of the present invention, the picture annotation information includes a picture annotation type and an annotation data type, the annotation type is divided into a scene matching type and a non-scene matching type, the annotation data type is divided into a text type and a non-text annotation type, and the scene text is filtered out from each target annotation content text according to the to-be-matched feature information of the empirical screenshot of each target indicator, including:

[0016] Select the target index empirical screenshots with the picture annotation type being the scene matching type and the annotation data type being the text type from the set of target index empirical screenshots;

[0017] Classify the target annotation content text of the selected target index empirical screenshots, and extract the text content features from the text classification set with the largest number of texts as the scene text.

[0018] According to an implementable manner of the first aspect of the present invention, the scene matching of each target index empirical screenshot in the set of target index empirical screenshots according to the scene text includes:

[0019] Set the preset same scene condition as: the picture annotation type of the target index empirical screenshot is the scene matching type, the annotation data type of the target index empirical screenshot is the text type, and the maximum value of the picture acquisition time difference of the target index empirical screenshot set to which it belongs is not greater than the preset time threshold.

[0020] According to an implementable manner of the first aspect of the present invention, the method further includes:

[0021] Perform a review desensitization data detection on the target index empirical screenshots determined to belong to the same scene, and perform a desensitization process on the corresponding target index empirical screenshots according to the detected review desensitization data.

[0022] According to an implementable manner of the first aspect of the present invention, the performing a review desensitization data detection on the target index empirical screenshots determined to belong to the same scene, and performing a desensitization process on the corresponding target index empirical screenshots according to the detected review desensitization data includes:

[0023] Collect review desensitization data, where the review desensitization data includes standard review data and standard desensitization data;

[0024] Match the target annotation content text of the target index empirical screenshots belonging to the same scene with the review desensitization data, calculate the review annotation information of the matched review text according to the preset review annotation rules, and calculate the desensitization annotation information of the matched desensitization text according to the preset desensitization annotation rules;

[0025] Perform data annotation drawing on the corresponding target index empirical screenshots according to the review annotation information and the desensitization annotation information.

[0026] According to an implementable manner of the first aspect of the present invention, the performing a review desensitization data detection on the target index empirical screenshots determined to belong to the same scene, and performing a desensitization process on the corresponding target index empirical screenshots according to the detected review desensitization data further includes:

[0027] The preset review annotation rule is set as follows: the annotation coordinate area corresponding to the review text is used as the data annotation area, and the dotted hollow rectangle is used as the data annotation graphic type;

[0028] And / or, the preset desensitizing annotation rule is set as: dividing the annotation coordinate area corresponding to the desensitized text into three sub-areas, taking the middle sub-area of the three sub-areas as the data annotation area, and taking the solid hollow rectangle as the data annotation graphic type.

[0029] A second aspect of the present invention provides a device for processing screenshots of medical rating indicator evidence, comprising:

[0030] A picture set construction module, which is used to construct a corresponding target indicator empirical screenshot set by taking screenshots of the target medical software functional interface as target indicator empirical screenshots;

[0031] The image feature extraction module is used to perform similarity matching between each target indicator empirical screenshot in the target indicator empirical screenshot set and a medical software function interface screenshot carrying image feature information in the image feature information library, and obtain corresponding image feature information according to the matching result as the corresponding target image feature information; the image feature information includes image annotation information and image resolution;

[0032] An image scaling ratio calculation module is used to obtain the image resolution of the target indicator demonstration screenshot as the corresponding target image resolution, and calculate the image scaling ratio of the corresponding target indicator demonstration screenshot according to the target image resolution and the corresponding target image feature information;

[0033] A target annotation content text acquisition module is used to calculate the target image annotation information of the corresponding target indicator empirical screenshot according to the calculated image zoom ratio and the image annotation information in the corresponding target image feature information, identify the annotation coordinate area of the target image annotation information, and obtain the corresponding target annotation content text according to the annotation coordinate area;

[0034] A scene text screening module is used to use the target image annotation information and the target annotation content text as the feature information to be matched of the target indicator empirical screenshot, and to screen out the scene text from each target annotation content text according to the feature information to be matched of each target indicator empirical screenshot;

[0035] The image scene matching module is used to perform scene matching on each target indicator verification screenshot in the target indicator verification screenshot set according to the scene text. If the target image annotation information of the target indicator verification screenshot meets the preset same scene condition and the corresponding target annotation content text matches the scene text, it is determined that the corresponding target indicator verification screenshots belong to the same scene.

[0036] According to an implementable manner of the second aspect of the present invention, the picture annotation information includes annotation coordinates, and the target annotation content text acquisition module includes:

[0037] A target picture annotation information calculation unit, configured to use the product of the calculated picture scaling ratio and the corresponding annotation coordinates as the target annotation coordinates of the corresponding target index empirical screenshot.

[0038] According to an implementable manner of the second aspect of the present invention, the picture annotation information includes a picture annotation type and an annotation data type. The annotation type is divided into a scene matching type and a non-scene matching type. The scene text screening module includes:

[0039] A first screening unit, configured to screen target index empirical screenshots with a picture annotation type of the scene matching type and an annotation data type of the text type from the set of target index empirical screenshots;

[0040] A second screening unit, configured to classify the target annotation content text of the screened target index empirical screenshots, and extract text content features from the text classification set with the largest number of texts as the scene text.

[0041] According to an implementable manner of the second aspect of the present invention, the picture scene matching module includes:

[0042] A first setting unit, configured to set the preset same scene condition as: the picture annotation type of the target index empirical screenshot is the scene matching type, the annotation data type of the target index empirical screenshot is the text type, and the maximum value of the picture acquisition time difference of the target index empirical screenshot set does not exceed the preset time threshold.

[0043] According to an implementable manner of the second aspect of the present invention, the device further includes:

[0044] A picture desensitization review module, configured to review desensitization data detection for target index empirical screenshots determined to belong to the same scene, and perform desensitization processing on the corresponding target index empirical screenshots according to the detected review desensitization data.

[0045] According to an implementable manner of the second aspect of the present invention, the picture desensitization review module includes:

[0046] A data acquisition unit, configured to acquire review desensitization data, where the review desensitization data includes standard review data and standard desensitization data;

[0047] An annotation information calculation unit, configured to match the target annotation content text of the target index empirical screenshot belonging to the same scene with the reviewed desensitized data, calculate the review annotation information of the reviewed text matched according to a preset review annotation rule, and calculate the desensitization annotation information of the desensitized text matched according to a preset desensitization annotation rule;

[0048] An annotation processing unit, configured to perform data annotation drawing on the corresponding target index empirical screenshot according to the review annotation information and the desensitization annotation information.

[0049] According to an implementable manner of the second aspect of the present invention, the picture desensitization review module further includes:

[0050] A second setting unit, configured to set the preset review annotation rule as: using the annotation coordinate area corresponding to the reviewed text as the data annotation area, and using a dashed-line hollow rectangle as the data annotation graphic type;

[0051] And / or, a third setting unit, configured to set the preset desensitization annotation rule as: dividing the annotation coordinate area corresponding to the desensitized text into three sub-areas, using the middle sub-area of the three sub-areas as the data annotation area, and using a solid-line hollow rectangle as the data annotation graphic type.

[0052] The third aspect of the present invention provides a processing device for medical rating index empirical screenshots, including:

[0053] A memory, configured to store instructions; wherein, the instructions are used to implement the processing method for medical rating index empirical screenshots according to any one of the above implementable manners;

[0054] A processor, configured to execute the instructions in the memory.

[0055] The fourth aspect of the present invention provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, it implements the processing method for medical rating index empirical screenshots according to any one of the above implementable manners.

[0056] From the above technical solutions, it can be seen that the present invention has the following advantages:

[0057] The present invention performs similarity matching between the empirical screenshots of each target index and the screenshots of the medical software function interfaces carrying picture feature information in the picture feature information library to obtain the target picture feature information; calculates the picture scaling ratio according to the picture resolution of the empirical screenshot of the target index and the corresponding target picture feature information, calculates the target picture annotation information according to the calculated picture scaling ratio and the picture annotation information in the corresponding target picture feature information, and performs picture annotation area recognition, and finally obtains the corresponding target annotation content text; uses the target picture annotation information and the target annotation content text as the feature information to be matched of the empirical screenshot of the target index, and filters out the scenario text from each target annotation content text according to the feature information to be matched of each empirical screenshot of the target index; performs scenario matching on each empirical screenshot of the target index in the set of empirical screenshots of the target index according to the scenario text. If the target picture annotation information of the empirical screenshot of the target index meets the preset same-scenario condition and the corresponding target annotation content text matches the scenario text, it is determined that the corresponding empirical screenshot of the target index belongs to the same scenario; the present invention determines the core feature area of the empirical screenshot of the target index by combining picture annotation and image similarity matching, thereby obtaining the target annotation content text for the core feature area, avoiding text recognition for invalid areas, reducing the complexity of picture recognition, and the subsequent scenario matching method is simple and convenient, and can effectively improve the recognition accuracy and processing efficiency as a whole. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the drawings in the following description are only some embodiments of the present invention, and for those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0059] Figure 1 It is a flowchart of a method for processing empirical screenshots of medical rating indicators provided by an optional embodiment of the present invention;

[0060] Figure 2 It is a flowchart of a method for processing empirical screenshots of medical rating indicators provided by another optional embodiment of the present invention;

[0061] Figure 3 It is a structural connection block diagram of a device for processing empirical screenshots of medical rating indicators provided by an optional embodiment of the present invention;

[0062] Figure 4 It is a structural connection block diagram of a device for processing empirical screenshots of medical rating indicators provided by another optional embodiment of the present invention.

[0063] Reference Signs:

[0064] 1 - Picture set construction module; 2 - Picture feature extraction module; 3 - Picture scaling ratio calculation module; 4 - Target annotation content text acquisition module; 5 - Scene text screening module; 6 - Picture scene matching module; 7 - Picture desensitization review module. Detailed implementation manners

[0065] The embodiments of the present invention provide a method, device and equipment for processing index empirical screenshots for medical rating, which are used to solve the technical problems of high recognition error rate and low efficiency existing in the existing image intelligent matching technology when processing index empirical screenshots.

[0066] In order to make the invention purpose, features and advantages of the present invention more obvious and understandable, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the embodiments described below are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without making creative efforts belong to the scope of protection of the present invention.

[0067] The present invention provides a method for processing index empirical screenshots for medical rating.

[0068] Please refer to Figure 1 , Figure 1 which shows a flowchart of a method for processing index empirical screenshots for medical rating provided by an embodiment of the present invention.

[0069] A method for processing index empirical screenshots for medical rating provided by an embodiment of the present invention includes steps S1 - S6.

[0070] Step S1, using the screenshots of the target medical software function interface as the target index empirical screenshots, and constructing a corresponding target index empirical screenshot set.

[0071] The target index empirical screenshots in this embodiment refer to the index empirical screenshots to be matched.

[0072] Step S2, performing similarity matching on each target index empirical screenshot in the target index empirical screenshot set with the screenshots of the medical software function interface carrying picture feature information in the picture feature information library, and obtaining the corresponding picture feature information according to the obtained matching result as the corresponding target picture feature information; the picture feature information includes picture annotation information and picture resolution.

[0073] The picture feature information library is a pre - constructed database, which stores multiple screenshots of the medical software function interface carrying picture feature information.

[0074] To build an image feature information library, information features of medical software interface images are collected, that is, the software interface is screenshot manually, and each screenshot is processed with manual annotation to obtain the image annotation information of each screenshot. The image annotation information may include annotation coordinates, image annotation types, and annotation data types. At the same time, the resolution of each screenshot is identified, and the screenshot resolution and image annotation information are used as image feature information and stored in the corresponding position of the medical software function interface screenshot in the image feature information library.

[0075] Among them, the annotation is based on a rectangular box. The annotation coordinates can be represented by (x, y, width, height), where the lower left corner of the image is used as the reference point, x represents the x coordinate of the upper left corner of the annotation rectangular box, y represents the y coordinate of the upper left corner of the annotation rectangular box, width represents the width of the annotation rectangular box, and height represents the height of the annotation rectangular box.

[0076] The image annotation types are divided into scene matching types and non-scene matching types. The non-scene matching types include desensitization types. Among them, the images of the scene matching type are used for subsequent image scene matching and image review, and the desensitization type images can be used for subsequent desensitization processing of images.

[0077] The annotation data types can be divided into text types and non-text annotation types. The non-text annotation types include time types. The text type images can be used for subsequent image scene matching, and the time type images can be used for calculating the time difference of subsequent image acquisition.

[0078] In an implementable way, the similarity matching is performed between each target index empirical screenshot in the target index empirical screenshot set and the medical software function interface screenshot carrying image feature information in the image feature information library, including:

[0079] The image similarity matching is performed according to the following formula:

[0080]

[0081] In the formula, SSIM(A, B) represents the similarity between images A and B, μ A represents the average brightness of image A, μ B represents the average brightness of image B, σ A represents the standard deviation of the brightness of image A, σ B represents the standard deviation of the brightness of image B, σ AB represents the brightness covariance between image A and image B, c1, c2, and c3 are constants, usually taking c1 = k1L, c2 = k2L, c3 = k3L, L is the number of gray levels of the image, k1 = k2 = 0.001, k3 = 0.03.

[0082] In this embodiment, the similarity between two images is calculated based on the structure, contrast, and brightness of the images. If the value of SSIM(A,B) tends to 1, it indicates that the two images are similar. The corresponding similarity threshold can be set according to the actual situation to determine whether the two images are similar.

[0083] It should be noted that the target empirical screenshots in the target indicator empirical screenshot set can also be matched for similarity with the medical software function interface screenshots carrying picture feature information in the picture feature information library according to other existing picture similarity matching methods.

[0084] The corresponding picture feature information is obtained according to the obtained matching result as the corresponding target picture feature information, that is, the picture feature information of the matched medical software function interface screenshot is used as the corresponding target picture feature information. In a realizable manner, if no matching medical software function interface screenshot can be found in the picture feature information library, information for prompting that there is no matching medical software function interface screenshot for the current target indicator empirical screenshot is output, so as to facilitate the relevant personnel to manually process the target indicator empirical screenshot.

[0085] Step S3, obtain the picture resolution of the target indicator empirical screenshot as the corresponding target picture resolution, and calculate the picture scaling ratio of the corresponding target indicator empirical screenshot according to the target picture resolution and the corresponding target picture feature information.

[0086] In a realizable manner, the picture scaling ratio of the corresponding target indicator empirical screenshot is calculated according to the following formula:

[0087]

[0088] In the formula, n represents the picture scaling ratio, (width1, height1) is the target picture resolution, and (width2, height2) is the picture resolution in the corresponding target picture feature information of the target indicator empirical screenshot.

[0089] In another realizable manner, the corresponding correction coefficient can be set according to the actual situation to correct the picture scaling ratio calculated by the previous realizable manner, and the corrected value is used as the final picture scaling ratio to improve the accuracy of the picture scaling ratio calculation.

[0090] Step S4, calculate the target picture annotation information of the corresponding target indicator empirical screenshot according to the calculated picture scaling ratio and the picture annotation information in the corresponding target picture feature information, identify the annotation coordinate area of the target picture annotation information, and obtain the corresponding target annotation content text according to the annotation coordinate area.

[0091] In one achievable manner, the image annotation information includes annotation coordinates, and the target image annotation information of the corresponding target indicator empirical screenshot is calculated according to the calculated image scaling ratio and the image annotation information in the corresponding target image feature information, including:

[0092] The product of the calculated image scaling ratio and the corresponding annotation coordinates is used as the target annotation coordinates of the corresponding target indicator empirical screenshot.

[0093] The specific calculation method is as follows:

[0094] ρ=(x,y,width,height)×n

[0095] Where ρ represents the target annotation coordinates of the target indicator empirical screenshot, (x, y, width, height) is the annotation coordinates in the corresponding target image feature information, and n is the calculated image scaling ratio.

[0096] In the above embodiment of the present invention, the core feature area of the target indicator empirical screenshot is determined by combining picture annotation with image similarity matching, so as to obtain the target annotation content text for the core feature area, avoid text recognition in invalid areas, reduce the complexity of picture recognition, thereby shortening the recognition time and improving the recognition success rate.

[0097] Step S5, using the target image annotation information and the target annotation content text as the feature information to be matched of the target indicator verification screenshot, and filtering out the scene text from each target annotation content text according to the feature information to be matched of each target indicator verification screenshot.

[0098] In one achievable manner, the image annotation information includes an image annotation type and an annotation data type, the annotation type is divided into a scene matching type and a non-scene matching type, the annotation data type is divided into a text type and a non-text annotation type, and the scene text is screened out from each target annotation content text according to the to-be-matched feature information of each target indicator empirical screenshot, including:

[0099] Filtering target indicator empirical screenshots whose image annotation type is a scene matching type and whose annotation data type is a text type from the target indicator empirical screenshot set;

[0100] The target annotation content text of the screened target indicator empirical screenshots is classified, and the text content features are extracted from the text classification set with the largest number of texts as the scene text.

[0101] When performing the first - step screening, specifically, each target - index empirical screenshot in the set of target - index empirical screenshots can be classified according to the picture annotation type, and then the target - index empirical screenshots belonging to the scenario - matching type can be further classified according to the annotation data type to obtain the target - index empirical screenshots belonging to the text type.

[0102] When performing the second - step screening, specifically, feature keywords can be extracted from the target - annotation content text, and the extracted feature keywords are used as the identifiers of the target - annotation content text. Then, based on the identifiers, the target - annotation content texts of each target - index empirical screenshot are classified. The feature keywords can be set according to the actual situation, such as keywords representing operator information, keywords representing operation time, etc.

[0103] Correspondingly, when extracting the text - content features as the scenario text from the text - classification set with the largest number of texts, the identifier of the text - classification set can be used as the text - content feature from the text - classification set with the largest number of texts.

[0104] In other embodiments, the target - annotation content text with the fewest text - content keywords in the text - classification set can also be used as the scenario text.

[0105] In other embodiments, when performing the second - step screening, the target - annotation content text can also be pre - processed to filter out meaningless keywords, so as to improve the screening performance.

[0106] It should be noted that the target - annotation content texts of the selected target - index empirical screenshots can also be classified according to the existing text - classification methods.

[0107] Step S6: Perform scenario matching on each target - index empirical screenshot in the set of target - index empirical screenshots according to the scenario text. If the target - picture annotation information of the target - index empirical screenshot meets the preset same - scenario conditions and the corresponding target - annotation content text matches the scenario text, it is determined that the corresponding target - index empirical screenshot belongs to the same scenario.

[0108] In an implementable manner, the preset same - scenario conditions are set as follows: the picture annotation type of the target - index empirical screenshot is the scenario - matching type, the annotation data type of the target - index empirical screenshot is the text type, and the maximum value of the picture - acquisition time difference of the target - index empirical screenshot set does not exceed the preset time threshold.

[0109] Among them, the maximum value of the picture - acquisition time difference of the corresponding target - index empirical screenshot set can be calculated based on the picture - acquisition time of the target - index empirical screenshot with the annotation data type of the time type. The picture - acquisition time can be obtained from the corresponding picture annotation information.

[0110] Since currently when medical institutions conduct the grading evaluation of the application level of electronic medical record devices, the function interfaces in the index empirical screenshots must be recent software function interfaces, usually within 3 months, the preset time threshold can be set to 3 months here.

[0111] In an implementable manner, as Figure 2 shown, on the basis of the method shown in Figure 1 the method further includes:

[0112] Step S7, conduct a review of desensitization data detection on the target index empirical screenshots determined to belong to the same scenario, and perform desensitization processing on the corresponding target index empirical screenshots according to the detected review desensitization data.

[0113] In an implementable manner, the conducting a review of desensitization data detection on the target index empirical screenshots determined to belong to the same scenario, and performing desensitization processing on the corresponding target index empirical screenshots according to the detected review desensitization data includes:

[0114] Collect review desensitization data, where the review desensitization data includes standard review data and standard desensitization data;

[0115] Match the target annotation content text of the target index empirical screenshots belonging to the same scenario with the review desensitization data, calculate the review annotation information of the reviewed text matched according to the preset review annotation rules, and calculate the desensitization annotation information of the desensitized text matched according to the preset desensitization annotation rules;

[0116] Perform data annotation drawing on the corresponding target index empirical screenshots according to the review annotation information and the desensitization annotation information.

[0117] There are various ways to collect review desensitization data. For example, sensitive data that needs to be reviewed and desensitized can be manually entered, an Excel table of review desensitization data can be imported, an SQL script for review desensitization data and the data source information for executing the SQL can be provided, etc.

[0118] In an implementable manner, the preset review annotation rules can be set as: using the annotation coordinate area corresponding to the reviewed text as the data annotation area, and using a dotted-line hollow rectangle as the data annotation graphic type. The preset desensitization annotation rules can be set as: dividing the annotation coordinate area corresponding to the desensitized text into three sub-areas, using the middle sub-area of the three sub-areas as the data annotation area, and using a solid-line hollow rectangle as the data annotation graphic type.

[0119] When performing data annotation drawing on the corresponding target index empirical screenshot according to the review annotation information and the desensitized annotation information, a dotted-line hollow rectangle can be drawn in the data annotation area corresponding to the review text, and a solid-line hollow rectangle can be drawn in the data annotation area corresponding to the desensitized text.

[0120] It should be noted that other preset review annotation rules and preset desensitized annotation rules can also be set according to the actual situation. For example, other data annotation graphic types or the determination principle of the annotation coordinate area can be set.

[0121] In other embodiments, encryption processing, deletion or replacement processing can also be directly performed on the data annotation area corresponding to the review text and the data annotation area corresponding to the desensitized text.

[0122] The present invention also provides a processing device for medical rating index empirical screenshots, and this device can be used to execute the processing method for medical rating index empirical screenshots described in any one of the above embodiments of the present invention.

[0123] Please refer to Figure 3 , Figure 3 which shows a structural connection block diagram of a processing device for medical rating index empirical screenshots provided by an embodiment of the present invention.

[0124] A processing device for medical rating index empirical screenshots provided by an embodiment of the present invention includes:

[0125] A picture set construction module 1, which is used to construct a corresponding set of target index empirical screenshots with the screenshots of the target medical software function interface as the target index empirical screenshots;

[0126] A picture feature extraction module 2, which is used to perform similarity matching between each target index empirical screenshot in the set of target index empirical screenshots and the medical software function interface screenshots carrying picture feature information in the picture feature information library, and obtain the corresponding picture feature information as the corresponding target picture feature information according to the obtained matching result; the picture feature information includes picture annotation information and picture resolution;

[0127] A picture scaling ratio calculation module 3, which is used to obtain the picture resolution of the target index empirical screenshot as the corresponding target picture resolution, and calculate the picture scaling ratio of the corresponding target index empirical screenshot according to the target picture resolution and the corresponding target picture feature information;

[0128] The target annotation content text acquisition module 4 is used to calculate the target image annotation information of the corresponding target index empirical screenshot according to the calculated image scaling ratio and the image annotation information in the corresponding target image feature information, identify the annotation coordinate area of the target image annotation information, and obtain the corresponding target annotation content text according to the annotation coordinate area;

[0129] The scene text screening module 5 is used to use the target image annotation information and the target annotation content text as the to-be-matched feature information of the target index empirical screenshot, and screen out the scene text from each target annotation content text according to the to-be-matched feature information of each target index empirical screenshot;

[0130] The image scene matching module 6 is used to perform scene matching on each target index empirical screenshot in the target index empirical screenshot set according to the scene text. If the target image annotation information of the target index empirical screenshot meets the preset same-scene condition and the corresponding target annotation content text matches the scene text, it is determined that the corresponding target index empirical screenshot belongs to the same scene.

[0131] In an implementable manner, the image annotation information includes annotation coordinates, and the target annotation content text acquisition module 4 includes:

[0132] The target image annotation information calculation unit is used to use the product of the calculated image scaling ratio and the corresponding annotation coordinates as the target annotation coordinates of the corresponding target index empirical screenshot.

[0133] In an implementable manner, the image annotation information includes an image annotation type and an annotation data type. The annotation type is divided into a scene matching type and a non-scene matching type. The scene text screening module 5 includes:

[0134] The first screening unit is used to screen out the target index empirical screenshots from the target index empirical screenshot set whose image annotation type is the scene matching type and whose annotation data type is the text type;

[0135] The second screening unit is used to classify the target annotation content text of the screened target index empirical screenshots, and extract the text content features from the text classification set with the largest number of texts as the scene text.

[0136] In an implementable manner, the image scene matching module 6 includes:

[0137] The first setting unit is used to set the preset same-scene condition as: the image annotation type of the target index empirical screenshot is the scene matching type, the annotation data type of the target index empirical screenshot is the text type, and the maximum value of the image acquisition time difference of the target index empirical screenshot set to which it belongs is not greater than the preset time threshold.

[0138] In an implementable manner, as Figure 4 shown, on the basis of the device shown Figure 3 , the device further includes:

[0139] A picture desensitization review module 7, configured to perform review desensitization data detection on the empirical screenshots of the target indicators determined to belong to the same scene, and desensitize the corresponding empirical screenshots of the target indicators according to the detected review desensitization data.

[0140] In an implementable manner, the picture desensitization review module 7 includes:

[0141] A data acquisition unit, configured to acquire review desensitization data, where the review desensitization data includes standard review data and standard desensitization data;

[0142] A labeling information calculation unit, configured to match the target labeling content text of the empirical screenshots of the target indicators belonging to the same scene with the review desensitization data, calculate the review labeling information of the matched review text according to the preset review labeling rules, and calculate the desensitization labeling information of the matched desensitization text according to the preset desensitization labeling rules;

[0143] A labeling processing unit, configured to perform data labeling drawing on the corresponding empirical screenshots of the target indicators according to the review labeling information and the desensitization labeling information.

[0144] In an implementable manner, the picture desensitization review module 7 further includes:

[0145] A second setting unit, configured to set the preset review labeling rules as: using the labeling coordinate area corresponding to the review text as the data labeling area, and using a dashed hollow rectangle as the data labeling graphic type;

[0146] And / or, a third setting unit, configured to set the preset desensitization labeling rules as: dividing the labeling coordinate area corresponding to the desensitization text into three sub-areas, using the middle sub-area of the three sub-areas as the data labeling area, and using a solid hollow rectangle as the data labeling graphic type.

[0147] The present invention further provides a processing device for empirical screenshots of medical rating indicators, including:

[0148] A memory, configured to store instructions; wherein, the instructions are used to implement the processing method for empirical screenshots of medical rating indicators described in any one of the above embodiments;

[0149] A processor, configured to execute the instructions in the memory.

[0150] The present invention also provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, it implements the processing method for the empirical screenshots of medical rating indicators as described in any one of the above embodiments.

[0151] It should be noted that the processing method, device and equipment for the empirical screenshots of medical rating indicators in the present application can be applied not only to the medical rating scenarios described in the background technology of the present application, but also to other scenarios that require image matching processing.

[0152] Those skilled in the art can clearly understand that for the convenience and simplicity of description, the specific working processes of the above-described devices, equipment and modules can refer to the corresponding processes in the foregoing method embodiments, and the specific beneficial effects of the above-described devices, equipment and modules can refer to the corresponding beneficial effects in the foregoing method embodiments, which will not be elaborated herein.

[0153] In several embodiments provided by the present application, it should be understood that the disclosed devices, equipment and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the division of the modules is only a logical function division. In actual implementation, there may be other division methods. For example, multiple modules or components can be combined or integrated into another device, or some features can be ignored or not executed. Another point is that the displayed or discussed couplings or direct couplings or communication connections to each other can be through some interfaces. The indirect couplings or communication connections of the devices or modules can be in electrical, mechanical or other forms.

[0154] The modules described as separate components may or may not be physically separated. The components shown as modules may or may not be physical modules, that is, they can be located in one place, or distributed to multiple network modules. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0155] In addition, in each embodiment of the present invention, the functional modules can be integrated into one processing module, or each module can exist physically alone, or two or more modules can be integrated into one module. The above integrated modules can be implemented in the form of hardware or in the form of software functional modules.

[0156] When the integrated module is implemented in the form of a software functional module and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of this technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in various embodiments of the present invention. The aforementioned storage medium includes: various media that can store program codes, such as USB flash drives, mobile hard disks, read-only memories (ROM, Read-Only Memory), random access memories (RAM, Random Access Memory), magnetic disks, or optical discs.

[0157] As described above, the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions described in the foregoing embodiments or perform equivalent replacements for some of the technical features; and these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of various embodiments of the present invention.

Claims

1. A method for processing empirical screenshots of medical rating indicators, characterized in that, include: Taking screenshots of the functional interface of the target medical software as the target indicator empirical screenshots, a corresponding target indicator empirical screenshot set is constructed; Performing similarity matching on each target indicator empirical screenshot in the target indicator empirical screenshot set and a medical software function interface screenshot carrying image feature information in the image feature information library, and obtaining corresponding image feature information as corresponding target image feature information according to the matching result; the image feature information includes image annotation information and image resolution; Obtaining a picture resolution of the target indicator demonstration screenshot as the corresponding target picture resolution, and calculating a picture scaling ratio of the corresponding target indicator demonstration screenshot according to the target picture resolution and the corresponding target picture feature information; Calculate the target image annotation information of the corresponding target indicator empirical screenshot according to the calculated image zoom ratio and the image annotation information in the corresponding target image feature information, identify the annotation coordinate area of the target image annotation information, and obtain the corresponding target annotation content text according to the annotation coordinate area; The target image annotation information and the target annotation content text are used as the feature information to be matched of the target indicator empirical screenshot, and the scene text is filtered out from each target annotation content text according to the feature information to be matched of each target indicator empirical screenshot; Perform scene matching on each target indicator demonstration screenshot in the target indicator demonstration screenshot set according to the scene text, and if the target image annotation information of the target indicator demonstration screenshot meets the preset same scene condition and the corresponding target annotation content text matches the scene text, then it is determined that the corresponding target indicator demonstration screenshots belong to the same scene; The image annotation information includes annotation coordinates, and the target image annotation information of the corresponding target indicator verification screenshot is calculated based on the calculated image zoom ratio and the image annotation information in the corresponding target image feature information, including: taking the product of the calculated image zoom ratio and the corresponding annotation coordinates as the target annotation coordinates of the corresponding target indicator verification screenshot.

2. The processing method for the empirical screenshot of medical rating indicators according to claim 1, wherein The image annotation information includes an image annotation type and an annotation data type, wherein the annotation type is divided into a scene matching type and a non-scene matching type, and the annotation data type is divided into a text type and a non-text annotation type. The scene text is filtered out from the annotation content text of each target according to the to-be-matched feature information of the empirical screenshot of each target indicator, including: Filtering target indicator empirical screenshots whose image annotation type is a scene matching type and whose annotation data type is a text type from the target indicator empirical screenshot set; The target annotation content text of the screened target indicator empirical screenshots is classified, and the text content features are extracted from the text classification set with the largest number of texts as the scene text.

3. The processing method for the empirical screenshot of medical rating indicators according to claim 2, characterized in that The performing scene matching on each target indicator demonstration screenshot in the target indicator demonstration screenshot set according to the scene text includes: The preset same scene condition is set as follows: the image annotation type of the target indicator verification screenshot is the scene matching type, the annotation data type of the target indicator verification screenshot is the text type, and the maximum value of the image collection time difference of the target indicator verification screenshot set is not greater than the preset time threshold.

4. The processing method of the empirical screenshot for medical rating according to claim 1, characterized in that The method further comprises: The target indicator empirical screenshots that are determined to belong to the same scenario are subject to review and desensitized data detection, and the corresponding target indicator empirical screenshots are desensitized based on the detected review and desensitized data.

5. The processing method of the empirical screenshot for medical rating according to claim 4, characterized in that, The step of performing review and desensitization data detection on the target indicator empirical screenshots determined to belong to the same scenario, and desensitizing the corresponding target indicator empirical screenshots according to the detected review and desensitization data, includes: Collecting review anonymized data, wherein the review anonymized data includes standard review data and standard anonymized data; Match the target annotation content text of the target indicator empirical screenshot belonging to the same scenario with the review desensitized data, calculate the review annotation information of the matched review text according to the preset review annotation rules, and calculate the desensitized annotation information of the matched desensitized text according to the preset desensitization annotation rules; Data annotation and drawing are performed on the corresponding target indicator empirical screenshots according to the review annotation information and the desensitized annotation information.

6. The processing method for the empirical screenshot of medical rating indicators according to claim 5, characterized in that The step of performing review and desensitization data detection on the target indicator evidence screenshots determined to belong to the same scenario, and desensitizing the corresponding target indicator evidence screenshots according to the detected review and desensitization data, further includes: The preset review annotation rule is set as follows: the annotation coordinate area corresponding to the review text is used as the data annotation area, and the dotted hollow rectangle is used as the data annotation graphic type; And / or, the preset desensitizing annotation rule is set as: dividing the annotation coordinate area corresponding to the desensitized text into three sub-areas, taking the middle sub-area of the three sub-areas as the data annotation area, and taking the solid hollow rectangle as the data annotation graphic type.

7. A processing device for medical rating index empirical screenshots, characterized in that, include: A picture set construction module, which is used to construct a corresponding target indicator empirical screenshot set by taking screenshots of the target medical software functional interface as target indicator empirical screenshots; The image feature extraction module is used to perform similarity matching between each target indicator empirical screenshot in the target indicator empirical screenshot set and a medical software function interface screenshot carrying image feature information in the image feature information library, and obtain corresponding image feature information according to the matching result as the corresponding target image feature information; the image feature information includes image annotation information and image resolution; An image scaling ratio calculation module is used to obtain the image resolution of the target indicator demonstration screenshot as the corresponding target image resolution, and calculate the image scaling ratio of the corresponding target indicator demonstration screenshot according to the target image resolution and the corresponding target image feature information; A target annotation content text acquisition module is used to calculate the target image annotation information of the corresponding target indicator empirical screenshot according to the calculated image zoom ratio and the image annotation information in the corresponding target image feature information, identify the annotation coordinate area of the target image annotation information, and obtain the corresponding target annotation content text according to the annotation coordinate area; A scene text screening module is used to use the target image annotation information and the target annotation content text as the feature information to be matched of the target indicator empirical screenshot, and to screen out the scene text from each target annotation content text according to the feature information to be matched of each target indicator empirical screenshot; The picture scene matching module is used to perform scene matching on each target index empirical screenshot in the target index empirical screenshot set according to the scene text. If the target picture annotation information of the target index empirical screenshot meets the preset same scene condition and the corresponding target annotation content text matches the scene text, it is determined that the corresponding target index empirical screenshot belongs to the same scene; The picture annotation information includes annotation coordinates, and the target annotation content text acquisition module includes: a target picture annotation information calculation unit, which is used to use the product of the calculated picture scaling ratio and the corresponding annotation coordinates as the target annotation coordinates of the corresponding target index empirical screenshot.

8. A processing device for empirical screenshots of medical rating indicators, characterized in that, It includes: A memory for storing instructions; wherein, the instructions are used to implement the processing method for the index empirical screenshot for medical rating as described in any one of claims 1-6; A processor for executing the instructions in the memory.

9. A computer-readable storage medium, characterized in that, A computer program is stored on the computer-readable storage medium, and when the computer program is executed by the processor, it implements the processing method for the index empirical screenshot for medical rating as described in any one of claims 1-6.

Images