Desensitization processing method, system and equipment for privacy big data and storage medium

Abstract

The invention discloses a privacy big data desensitization processing method, system and device and a storage medium, and relates to the technical field of medical data desensitization, and the method comprises the following steps: carrying out text erasing processing on an original CT image to obtain a first CT image, and dividing the first CT image into an original main body region and an original background region; performing grid division on the original main body area and the original background area to obtain a grid main body area and a grid background area; cT background desensitization processing is carried out on the grid background area based on gray level distribution of the grid main body area, and a desensitization background area is obtained; performing CT main body desensitization processing on the grid main body area to obtain a desensitized main body area, and obtaining a desensitized CT image; the method is used for solving the problem that when an existing medical data desensitization technology is used for desensitizing a CT image, privacy information on the image cannot be removed, a computer-aided diagnosis system cannot recognize and diagnose, and meanwhile it is guaranteed that observation and diagnosis of students on the image are not affected.

Description

Technical Field

[0001] The present invention relates to the technical field of medical data desensitization, and specifically to a desensitization processing method, system, device and storage medium for privacy big data. Background Art

[0002] Medical data desensitization technology refers to a series of technical means for processing personal sensitive information contained in medical data, so as to reduce the risk of data exposing personal privacy without affecting the data usage value, and protect patient privacy and data security.

[0003] When the existing medical data desensitization technology desensitizes CT images, it often desensitizes them through blurring, cropping and occlusion, or encryption processing; after desensitizing the CT images through these methods, although the key privacy information is hidden, it is extremely inconvenient to apply in medical teaching. Blurring will reduce the image clarity, and for medical teaching, it may cause students to be difficult to observe subtle lesion features. Cropping and occlusion may accidentally delete important feature information, which is not conducive to students' comprehensive learning; while the encrypted images need a specific decryption key to view. In a medical teaching environment, if multiple students use image data at the same time, the distribution and management of the key will become complicated, and the original CT images after decryption contain a lot of patient privacy information; and during the teaching process, students' access to CT images may rely on computer-aided diagnosis systems, which is not conducive to cultivating students' ability to independently interpret images and affects the improvement of their diagnostic skills and clinical thinking abilities; for example, in the patent application with the publication number CN109243584A, a management method and system for CT image desensitization data based on content uniqueness are disclosed. This solution only desensitizes the relevant data of CT images and does not desensitize the CT images themselves, and cannot prevent students from obtaining CT images in medical teaching and using computer-aided diagnosis; therefore, when the existing medical data desensitization technology desensitizes CT images, it cannot remove the privacy information on the images and make the computer-aided diagnosis system unable to recognize and diagnose, while ensuring that it does not affect students' observation and diagnosis of the images. Summary of the Invention

[0004] The present invention aims to solve at least one of the technical problems in the prior art to a certain extent. By performing text erasure processing on the original CT image, a first CT image is obtained and divided into an original main region and an original background region; and grid division is respectively performed to obtain a grid main region and a grid background region; based on the gray-scale distribution of the grid main region, CT background desensitization processing is performed on the grid background region to obtain a desensitized background region; CT main body desensitization processing is performed on the grid main region to obtain a desensitized main region and a desensitized CT image, so as to solve the problem that the existing medical data desensitization technology cannot remove the privacy information on the image and make the computer-aided diagnosis system unable to recognize and diagnose while ensuring that it does not affect students' observation and diagnosis of the image.

[0005] To achieve the above object, in a first aspect, the present application provides a desensitization processing method for privacy big data, including the following steps:

[0006] Perform text erasure processing on the original CT image to obtain a first CT image, and divide it into an original main region and an original background region;

[0007] Perform grid division on the original main region and the original background region respectively to obtain a grid main region and a grid background region;

[0008] Based on the gray-scale distribution of the grid main region, perform CT background desensitization processing on the grid background region to obtain a desensitized background region;

[0009] Perform CT main body desensitization processing on the grid main region to obtain a desensitized main region and a desensitized CT image.

[0010] Further, performing text erasure processing on the original CT image to obtain a first CT image and dividing it into an original main region and an original background region includes the following sub-steps:

[0011] For any text in the original CT image, denoted as the first text, obtain the horizontal and vertical coordinates of all the pixel points contained in the first text in the original CT image, and arrange all the horizontal coordinates and all the vertical coordinates in ascending order, denoted as the horizontal coordinate sequence and the vertical coordinate sequence respectively;

[0012] Obtain the minimum value and the maximum value in the horizontal coordinate sequence, denoted as XX and DX in sequence; obtain the minimum value and the maximum value in the vertical coordinate sequence, denoted as XY and DY in sequence;

[0013] In the original CT image, pixel points with coordinates [XX, XY], [DX, XY], [DX, DY], and [XX, DY] are sequentially connected, and the enclosed closed area is denoted as the text erasure area. The gray values of all pixel points within the text erasure area are changed to K0; the above process is repeated for all texts in the original CT image, and after completion, the first CT image is obtained.

[0014] The image area corresponding to the human tissue in the first CT image is divided, denoted as the original main area, and the image area in the first CT image excluding the original main area and the text erasure area is denoted as the original background area.

[0015] Furthermore, the original main area and the original background area are respectively subjected to grid division to obtain the grid main area and the grid background area, including the following sub-steps:

[0016] Set the size of the first division grid to a*a, and the size of the second division grid to b*b;

[0017] Use the first division grid to divide the original background area, and denote the divided grid area as the background grid, and mark it in the order from left to right and from top to bottom. For the coordinates of any background grid, denote it as BW(x, y), where x represents the column number where the background grid is located, and y represents the row number where the background grid is located; denote the original background area after completing the grid division as the grid background area;

[0018] Use the second division grid to divide the original main area, and denote the divided grid area as the main grid, and mark it in the order from left to right and from top to bottom. For the coordinates of any main grid, denote it as RW(i, j), where i represents the column number where the main grid is located, and j represents the row number where the main grid is located; denote the main background area after completing the grid division as the grid main area.

[0019] Furthermore, based on the gray distribution of the grid main area, perform CT background desensitization processing on the grid background area to obtain the desensitized background area, including the following sub-steps:

[0020] For any background grid in the grid background area, if it is adjacent to any main grid, it is marked as a background adjacent grid, and the adjacent main grid is marked as a main adjacent grid, otherwise it is marked as a background separated grid; repeat to obtain all background separated grids, obtain the total number B0 of background separated grids, and randomly select B0*k1% of the background separated grids and denote them as the first background separated grids, and then randomly select B0*k2% of the background separated grids from the remaining background separated grids and denote them as the second background separated grids, where k1 and k2 are set proportionality coefficients;

[0021] Randomly select B0 * k1% of the main grids, denoted as reference main grids. For any one of the reference main grids, denoted as the first reference grid, arrange the gray values corresponding to all the pixel points in the first reference grid in ascending order, denoted as the reference gray sequence; obtain the average value of the smallest p1% of the gray values in the reference gray sequence, denoted as the first reference value, and obtain the average value of the largest p1% of the gray values in the reference gray sequence, denoted as the second reference value. Then obtain the average value of the reference gray sequence, denoted as the third reference value. Randomly select one from the first reference value, the second reference value, and the third reference value and denote it as the reference gray value of the corresponding reference main grid; repeat to obtain the reference gray values of all the reference main grids, denoted as the reference gray set.

[0022] Further, performing CT background desensitization processing on the grid background area based on the gray distribution of the grid main area to obtain the desensitized background area further includes the following sub-steps:

[0023] Perform first background desensitization processing on all the first background-separated grids, including: taking any one of the first background-separated grids, denoted as the desensitized first grid, and obtaining the pixel point at the geometric center of the desensitized first grid, denoted as the central pixel point; draw a circle with the central pixel point as the center and 1 / 3 * a as the radius, denoted as the first desensitized area; randomly select e1 pixel points from the first desensitized area, denoted as the first pixel points, and draw circles with the first pixel points as the centers and 1 / 6 * a as the radii respectively, denoted as the second desensitized areas; randomly select a reference gray value from the reference gray set, denoted as the desensitized gray value; change the gray values of the pixel points in the first desensitized area and all the second desensitized areas to the desensitized gray value;

[0024] Perform second background desensitization processing on all the second background-separated grids, including: setting a first random change value C1, where C1 = -1, 0, or 1; for any one of the second background-separated grids, denoted as the desensitized second grid, when C1 = -1, reduce the gray values of all the pixel points in the second background-separated grid by M1; when C1 = 1, increase the gray values of all the pixel points in the second background-separated grid by M1; when C1 = 0, no processing is performed;

[0025] Mark the grid background area that has completed CT background desensitization processing as the desensitized background area.

[0026] Further, performing CT main body desensitization processing on the grid main area to obtain the desensitized main area, and obtaining the desensitized CT image further includes the following sub-steps:

[0027] Mark the main grids in the main grid area that are not marked as adjacent grids of the main body as separated grids of the main body; obtain the total number B1 of the separated grids of the main body, and randomly select B1*k1% of the separated grids of the main body as the first separated grids of the main body, and then randomly select B0*k2% of the separated grids of the main body from the remaining background separated grids as the second separated grids of the main body;

[0028] Set the second random variation value C2, C2 = -1 or 1; perform the first main body desensitization processing on all the first separated grids of the main body, including: mark any first separated grid of the main body as the first grid of the main body, obtain the geometric center of the first grid of the main body, denoted as the first main body center; draw a circle with the first main body center as the center and k3*2 / b as the radius, denoted as the first main body circular area; randomly select k4% of the pixel points from all the pixel points included in the first main body circular area, denoted as the first desensitized pixel points, and randomly take a value for C2 as the marking value of all the first desensitized pixel points; where k3 and k4 are set proportionality coefficients, 0 < {k3, k4} < 1;

[0029] Set the gray scale adjustment range [V1, V2], and adjust the corresponding gray scale value for any first desensitized pixel point according to the gray scale adjustment formula. The gray scale adjustment formula is as follows: where AF is the gray scale value of the first desensitized pixel point before adjustment; BF is the gray scale value of the first desensitized pixel point after adjustment; repeat the adjustment for all the first desensitized pixel points.

[0030] Furthermore, perform CT main body desensitization processing on the main grid area to obtain a desensitized main body area, and obtaining a desensitized CT image also includes the following sub-steps:

[0031] Then perform the second main body desensitization processing on all the second separated grids of the main body, including: mark any second separated grid of the main body as the second grid of the main body, obtain the geometric center of the second grid of the main body, denoted as the second main body center; draw a circle with the second main body center as the center and k5*2 / b as the radius, denoted as the second main body circular area; then draw a circle with the second main body center as the center and k6*2 / b as the radius, denoted as the third main body circular area, where k5 and k6 are set proportionality coefficients, 1 > k5 > k6 > 0; mark the annular area enclosed by the boundary of the second main body circular area and the boundary of the third main body circular area as the first annular area;

[0032] The first annular region is evenly divided into k7 fan-shaped regions, denoted as the first fan-shaped regions. k8 pixel points are randomly selected from each of the first fan-shaped regions, denoted as desensitized edge pixel points. A desensitized edge pixel point is randomly selected as the starting edge pixel point, and the desensitized edge pixel point that is the closest to the starting edge pixel point and is located in the same first fan-shaped region as the starting edge pixel point and has not been connected is connected. Then, the connected desensitized edge pixel point is used as the new starting edge pixel point, and the next starting edge pixel point is continued to be connected. If all the desensitized edge pixel points in the first fan-shaped region have been connected, then the desensitized edge pixel point that is the closest in the adjacent first fan-shaped region is selected in the clockwise direction for connection until all the desensitized edge pixel points are connected, and the last connected desensitized edge pixel point is connected to the first starting edge pixel point. The formed closed region is denoted as the first main desensitized region.

[0033] All the pixel points within the first main desensitized region are denoted as the second desensitized pixel points, and then a random value is taken for C2 as the marking value for all the second desensitized pixel points. Then, the gray values of the second desensitized pixel points are adjusted respectively through the gray adjustment formula.

[0034] The grid main region after the CT main body desensitization process is marked as the desensitized main region; and after completion, the corresponding original CT image is marked as the desensitized CT image.

[0035] In a second aspect, the present application provides a desensitization processing system for privacy big data, including an information division module, a grid division module, a background desensitization module, and a main body desensitization module;

[0036] The information division module includes an information erasure unit and a region division unit; the information erasure unit is used to perform text erasure processing on the original CT image to obtain the first CT image, and the region division unit is used to divide the first CT image into an original main region and an original background region;

[0037] The grid division module is used to perform grid division on the original main region and the original background region respectively to obtain a grid main region and a grid background region;

[0038] The background desensitization module performs CT background desensitization processing on the grid background region based on the gray distribution of the grid main region to obtain a desensitized background region;

[0039] The main body desensitization module performs CT main body desensitization processing on the grid main region to obtain a desensitized main region and obtains a desensitized CT image.

[0040] In a third aspect, the present application provides an electronic device, including a processor and a memory, where the memory stores computer-readable instructions, and when the computer-readable instructions are executed by the processor, the steps in the above method are run.

[0041] In a fourth aspect, the present application provides a storage medium, on which a computer program is stored. When the computer program is executed by a processor, it runs the steps in the above method.

[0042] Advantages of the present invention: By performing text erasure processing on the original CT image, a first CT image is obtained and divided into an original main body region and an original background region; the original main body region and the original background region are respectively divided into grids to obtain a grid main body region and a grid background region; based on the gray-scale distribution of the grid main body region, CT background desensitization processing is performed on the grid background region to obtain a desensitized background region; CT main body desensitization processing is performed on the grid main body region to obtain a desensitized main body region, and a desensitized CT image is obtained; it is possible to remove the privacy information on the image and make it unrecognizable by the computer-aided diagnosis system, while ensuring that it does not affect students' observation and diagnosis of the image.

[0043] By the gray-scale distribution of the main body region, a similar desensitized region is added to the background region. The advantage is that it can destroy the original regular gray-scale pattern of the background; it is difficult for the computer to recognize based on the conventional gray-scale features, effectively interfering with its automatic analysis and diagnosis of the image; for students, the background region does not directly participate in the diagnosis and will not cause interference, ensuring a smooth learning process; by dividing irregular regions in some of the main body regions and performing gray-scale adjustment under the condition that it is imperceptible to the human eye, the advantage is that it is difficult for the computer to match based on the established feature template, reducing the accuracy of its automatic diagnosis. For doctors' diagnosis and students' learning, since the gray-scale adjustment is within the range imperceptible to the human eye, the key feature information can still be clearly presented, without affecting students' judgment and learning, ensuring that the core value of the CT image in medical treatment and teaching is retained. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] Figure 1 It is a principle block diagram of the system of the present invention;

[0045] Figure 2 It is a step flow chart of the method of the present invention;

[0046] Figure 3 It is a schematic diagram of the first main body desensitization region of the present invention; [[ID=2,4]]

[0047] Figure 4 It is a schematic diagram of the structure of the electronic device of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0048] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0049] Example 1. Please refer to Figure 1 As shown, the present application provides a desensitization processing system for privacy big data, including an information division module, a grid division module, a background desensitization module, and a main body desensitization module;

[0050] The information division module includes an information erasure unit and a region division unit; the information erasure unit is used to perform text erasure processing on the original CT image to obtain a first CT image, and the region division unit is used to divide the first CT image into an original main body region and an original background region;

[0051] The information erasure unit is configured with an information erasure strategy, and the information erasure strategy includes: for any text in the original CT image, denoted as the first text, obtain the horizontal and vertical coordinates of all the pixel points contained in the first text in the original CT image, and arrange all the horizontal coordinates and all the vertical coordinates in ascending order, denoted as the horizontal coordinate sequence and the vertical coordinate sequence respectively; the text includes Chinese characters, letters, and symbols;

[0052] Obtain the minimum value and the maximum value in the horizontal coordinate sequence, denoted as XX and DX in sequence; obtain the minimum value and the maximum value in the vertical coordinate sequence, denoted as XY and DY in sequence;

[0053] In the original CT image, connect the pixel points with coordinates [XX, XY], [DX, XY], [DX, DY], and [XX, DY] in sequence, and denote the enclosed closed region as the text erasure region, and change the gray values of all the pixel points in the text erasure region to K0; repeat the processing of all the texts in the original CT image, and after completion, obtain the first CT image; in this embodiment, K0 is 0, that is, the text is blackened.

[0054] The region division unit is configured with a region division strategy, and the region division strategy includes: divide the image region corresponding to the human tissue in the first CT image, denoted as the original main body region, and denote the image region in the first CT image excluding the original main body region and the text erasure region as the original background region;

[0055] In the specific implementation process, the CT image usually contains patient basic information, scan parameter information, device information, and anatomical identification information, most of which may involve the patient's personal privacy and all need to be desensitized.

[0056] The mesh generation module is used to generate meshes for the original main region and the original background region respectively, obtaining the meshed main region and the meshed background region;

[0057] The mesh generation module is configured with a mesh generation strategy, which includes: setting the size of the first division mesh as a*a and the size of the second division mesh as b*b; since the CT image is rectangular, square meshes are generally selected for division, and their size units are generally measured by the number of pixels and can be set according to the actual application scenario. Generally, a > b;

[0058] Use the first division mesh to generate meshes for the original background region, record the divided mesh regions as background meshes, and mark them in the order from left to right and from top to bottom. The coordinates of any background mesh are denoted as BW(x, y), where x represents the column number where the background mesh is located, and y represents the row number where the background mesh is located; record the original background region after mesh generation as the meshed background region; for example, BW(3, 2) represents the background mesh located in the 3rd column and the 2nd row;

[0059] Use the second division mesh to generate meshes for the original main region, record the divided mesh regions as main meshes, and mark them in the order from left to right and from top to bottom. The coordinates of any main mesh are denoted as RW(i, j), where i represents the column number where the main mesh is located, and j represents the row number where the main mesh is located; record the main background region after mesh generation as the meshed main region; for example, RW(2, 4) represents the background mesh located in the 2nd column and the 4th row;

[0060] In the specific implementation process, generating meshes is for the convenience of subsequent processing. The coordinate markings of the background meshes and the main meshes should be marked separately. For example, the coordinate markings of the background meshes do not need to consider the main meshes.

[0061] The background desensitization module performs CT background desensitization processing on the meshed background region based on the gray level distribution of the meshed main region, obtaining the desensitized background region;

[0062] The background desensitization module is configured with a background desensitization strategy. The background desensitization strategy includes: for any background grid in the grid background area, if it is adjacent to any main grid, it is marked as a background adjacent grid, and the adjacent main grid is marked as a main adjacent grid; otherwise, it is marked as a background separated grid. Repeat to obtain all the background separated grids, obtain the total number B0 of the background separated grids, and randomly select B0*k1% of the background separated grids and record them as the first background separated grids. Then randomly select B0*k2% of the background separated grids from the remaining background separated grids and record them as the second background separated grids, where k1 and k2 are set proportionality coefficients. In this embodiment, k1% is 33% and k2% is 50%, that is, 1 / 3 of the background separated grids are selected as the first background separated grids, and then 1 / 2 of the background separated grids are selected as the second background separated grids.

[0063] Randomly select B0*k1% of the main grids and record them as reference main grids. For any reference main grid, recorded as the first reference grid, arrange the gray values corresponding to all the pixel points in the first reference grid in ascending order and record it as the reference gray sequence. Obtain the average value of the smallest p1% of the gray values in the reference gray sequence and record it as the first reference value, and obtain the average value of the largest p1% of the gray values in the reference gray sequence and record it as the second reference value. Then obtain the average value of the reference gray sequence and record it as the third reference value. Randomly select one from the first reference value, the second reference value, and the third reference value and record it as the reference gray value of the corresponding reference main grid. Repeat to obtain the reference gray values of all the reference main grids and record them as the reference gray set. In this embodiment, p1% is 30%.

[0064] The reference gray values obtained in this way comprehensively consider the distribution range and central tendency of the pixel gray levels in the main grid. This enables the selected reference gray values to more comprehensively represent the gray characteristics of the main grid, rather than relying solely on a single statistic, thus more accurately reflecting the gray information of the main area and providing a more reasonable basis for setting interfering areas in the background area later. And randomly selecting one as the reference gray value of the reference main grid introduces randomness and diversity. This random selection method makes the reference gray value of each reference main grid have a certain degree of uncertainty, avoiding fixed patterns or rules and increasing the difficulty of computer recognition.

[0065] Perform first background desensitization processing on all first background separated grids, including: Denote any one of the first background separated grids as the desensitized first grid, and obtain the pixel point at the geometric center of the desensitized first grid, denoted as the central pixel point; Draw a circle with the central pixel point as the center and 1 / 3*a as the radius, denoted as the first desensitization area; Randomly select e1 pixel points from the first desensitization area, denoted as the first pixel points, and draw circles with the first pixel points as the centers and 1 / 6*a as the radii respectively, denoted as the second desensitization areas; Randomly select a reference gray value from the reference gray value set, denoted as the desensitized gray value; Change the gray values of the pixel points in the first desensitization area and all the second desensitization areas to the desensitized gray value; In this embodiment, e1 is 4, and using 1 / 3*a and 1 / 6*a as the radii is to prevent the divided desensitization areas from exceeding the corresponding first background separated grids.

[0066] By randomly selecting points and drawing circles multiple times, the boundaries of the finally obtained areas form an irregular structure; This increases the complexity and diversity of the desensitization areas, enabling the computer to handle more details and variations when recognizing images, thereby enhancing the interference effect on computer vision algorithms, and making the shapes and grays of the desensitization areas within each first background separated grid different, greatly increasing the difficulty of computer analysis.

[0067] Perform second background desensitization processing on all second background separated grids, including: Set the first random change value C1, where C1 = -1, 0, or 1; For any one of the second background separated grids, denoted as the desensitized second grid, when C1 = -1, reduce the gray values of all pixel points in the second background separated grid by M1 respectively; When C1 = 1, increase the gray values of all pixel points in the second background separated grid by M1 respectively; When C1 = 0, no processing is performed; When processing each second background separated grid, C1 needs to be re - valued; In this embodiment, M1 is 5, and the value range is generally [3, 20].

[0068] Mark the grid background area that has completed CT background desensitization processing as the desensitized background area; Screen out the background adjacent grids without processing because they are adjacent to the main area, to avoid affecting the main area when processing them.

[0069] In the specific implementation process, the computer-aided diagnosis system for CT images often relies on gray-scale features for recognition; by migrating the gray-scale distribution features of the main body area to the background area to create a similar desensitized area, the original regular gray-scale pattern of the background can be disrupted; it is difficult for the computer to recognize based on the conventional gray-scale features of the background, effectively interfering with its automatic analysis and diagnosis of the image, and greatly reducing the risk of privacy information being improperly obtained or utilized by the computer; for students, the key to learning medical image diagnosis lies in mastering the features of the main body area; the background area does not directly participate in the diagnosis; although a similar desensitized area is added to the background area, when students observe the image, they will not be interfered by the background change and can still focus on the key information of the main body area, normally carry out diagnostic learning, ensure the smooth progress of the learning process, and accurately master knowledge.

[0070] The main body desensitization module performs CT main body desensitization processing on the grid main body area to obtain a desensitized main body area and a desensitized CT image;

[0071] The main body desensitization module is configured with a main body desensitization strategy, and the main body desensitization strategy includes: marking the main body grids in the grid main body area that are not marked as adjacent grids of the main body as separated grids of the main body; obtaining the total number B1 of the separated grids of the main body, and randomly selecting B1*k1% of the separated grids of the main body as the first separated grids of the main body, and then randomly selecting B0*k2% of the separated grids of the main body from the remaining background separated grids as the second separated grids of the main body; that is, selecting 1 / 3 of the separated grids of the main body as the first separated grids of the main body, and then selecting 1 / 3 of the separated grids of the main body as the second separated grids of the main body;

[0072] Set the second random change value C2, C2 = -1 or 1; perform the first main body desensitization processing on all the first separated grids of the main body respectively, including: marking any first separated grid of the main body as the first main body grid, obtaining the geometric center of the first main body grid, denoted as the first main body center; drawing a circle with the first main body center as the center and k3*2 / b as the radius, denoted as the first main body circular area; randomly selecting k4% of the pixel points from all the pixel points included in the first main body circular area, denoted as the first desensitized pixel points, and randomly taking a value for C2 as the marking value of all the first desensitized pixel points; where k3 and k4 are set proportionality coefficients, 0 < {k3, k4} < 1; in this embodiment, k = 0.8, k4% = 80%;

[0073] Set the gray-scale size adjustment range [V1, V2], and adjust the corresponding gray-scale value for any first desensitized pixel point according to the gray-scale adjustment formula. The gray-scale adjustment formula is as follows: where AF is the gray-scale value of the first desensitized pixel point before adjustment; BF is the gray-scale value of the first desensitized pixel point after adjustment; repeat the adjustment for all the first desensitized pixel points;

[0074] [V1, V2] can be set according to the actual application scenario. In this embodiment, [V1, V2] is [3, 12]. Because within the grayscale value range of 0 - 255, there is no fixed standard for the change in grayscale values that can be clearly perceived by the human eye. The human eye has the strongest resolution at the center position of the gray scale. Usually, within the grayscale value range of 80 - 160, a change of about 5 - 10 may be clearly perceived; while at both ends of the gray scale, such as in the ranges of 0 - 20 and 230 - 255, the human eye has weaker resolution, and a change of 1 - 20 or even more may be required to clearly perceive the change. When processing each grid separated by the first subject, C2 needs to be re - valued; that is, to ensure that within the circular area of the same first subject, the grayscale values either all increase or all decrease, just to prevent a significant grayscale contrast from occurring when processing two adjacent areas, one increasing and one decreasing, which may affect subsequent observation and diagnosis. The radius of the circular area of the first subject is set to k3 * 2 / b, also to prevent such a situation;

[0075] Then, perform second - subject desensitization processing on each grid separated by the second subject respectively, including: Denote any grid separated by the second subject as the second - subject grid, obtain the geometric center of the second - subject grid, denoted as the second - subject center; Draw a circle with the second - subject center as the center and k5 * 2 / b as the radius, denoted as the second - subject circular area; Then draw a circle with the second - subject center as the center and k6 * 2 / b as the radius, denoted as the third - subject circular area, where k5 and k6 are set proportionality coefficients, 1 > k5 > k6 > 0; Mark the annular area enclosed by the boundary of the second - subject circular area and the boundary of the third - subject circular area as the first annular area; Both k5 and k6 are less than 1, and k6 should have a certain gap with k5. In this embodiment, k5 = 0.8 and k6 = 0.5;

[0076] Please refer to Figure 3 As shown, evenly divide the first annular area into k7 sector areas, denoted as the first sector areas. Randomly select k8 pixel points from each first sector area, denoted as desensitization edge pixel points; Randomly select a desensitization edge pixel point as the starting edge pixel point, connect the desensitization edge pixel point that is the closest to the starting edge pixel point and is in the same first sector area as the starting edge pixel point and has not been connected, and then use the connected desensitization edge pixel point as the new starting edge pixel point to continue connecting the next starting edge pixel point; If all desensitization edge pixel points in the first sector area have been connected, then select the desensitization edge pixel point that is the closest in the adjacent first sector area in the clockwise direction for connection until all desensitization edge pixel points are connected, and connect the last connected desensitization edge pixel point to the first starting edge pixel point. Denote the formed closed area as the first - subject desensitization area; \n

[0077] That is, if there are still unconnected points in the current fan-shaped area, they will be connected first. Otherwise, the points closest to each other in adjacent fan-shaped areas will be connected in a clockwise direction, and finally, the head and tail will be connected to form a closed area. In this embodiment, k7 = 12 and k8 = 3. The purpose of doing this is to form a closed area with an irregular edge because the division of irregular areas can simulate the original shape and structure of the main area in the image, making it difficult for the computer to distinguish, increasing the difficulty of analyzing the image by computer vision algorithms. Moreover, compared with the desensitized areas of regular shapes, irregular areas are more difficult to be recognized by algorithms, improving the interference ability against computer algorithms.

[0078] All pixel points in the first main body desensitized area are denoted as second desensitized pixel points, and then a random value is taken for C2 as the marking value of all second desensitized pixel points. Then, the gray values of the second desensitized pixel points are adjusted respectively through the gray adjustment formula.

[0079] The grid main body area that has completed the CT main body desensitization process is marked as the desensitized main body area. And after completion, the corresponding original CT image is marked as the desensitized CT image. That is, the original CT image completes the text erasure process, grid division, CT background desensitization process, and CT main body desensitization process to obtain the desensitized CT image. In medical teaching, using the desensitized CT image to let students learn image diagnosis knowledge can not only ensure that students come into contact with real case images to cultivate diagnostic skills, but also avoid the exposure of patient information, and at the same time prevent the computer from automatically diagnosing and affecting the cultivation of students' independent thinking and judgment abilities.

[0080] In the specific implementation process, computer-aided diagnosis systems often rely on specific patterns and features in images for recognition and diagnosis. Dividing irregular areas and fine-tuning the gray levels in the main area can disrupt the conventional feature patterns used by the computer for recognition. For example, for an algorithm for recognizing liver lesions, the originally regular gray level distribution of the liver tissue is disrupted. The appearance of irregular areas and the gray level changes make it difficult for the computer to match according to the established feature templates, reducing the accuracy of its automatic diagnosis. For doctors' diagnosis and students' learning, although some main areas have been processed, since the gray level adjustment is within the range that is imperceptible to the human eye, the key lesion features and organ morphology can still be clearly presented, without affecting the judgment of diseases and learning, ensuring that the core value of medical images in medical treatment and teaching is retained.

[0081] Embodiment 2. Please refer to Figure 2 As shown, the present application provides a desensitization processing method for privacy big data, including the following steps:

[0082] Step S1: Perform a text erasure process on the original CT image to obtain a first CT image, and divide it into an original main area and an original background area. Step S1 includes the following sub-steps:

[0083] Step S101: For any text in the original CT image, denoted as the first text, obtain the horizontal and vertical coordinates of all the pixel points contained in the first text in the original CT image, and arrange all the horizontal coordinates and all the vertical coordinates in ascending order, denoted as the horizontal coordinate sequence and the vertical coordinate sequence respectively;

[0084] Step S102: Obtain the minimum value and the maximum value in the horizontal coordinate sequence, denoted as XX and DX in sequence; obtain the minimum value and the maximum value in the vertical coordinate sequence, denoted as XY and DY in sequence;

[0085] Step S103: Connect the pixel points with coordinates [XX, XY], [DX, XY], [DX, DY], and [XX, DY] in the original CT image in sequence, and denote the enclosed closed area as the text erasure area. Change the gray values of all the pixel points in the text erasure area to K0; Repeat the processing for all the texts in the original CT image. After completion, obtain the first CT image;

[0086] Step S104: Divide the image area corresponding to the human tissue in the first CT image, denoted as the original main area, and denote the image area in the first CT image excluding the original main area and the text erasure area as the original background area.

[0087] Step S2: Perform grid division on the original main area and the original background area respectively to obtain the grid main area and the grid background area; Step S2 includes the following sub-steps:

[0088] Step S201: Set the size of the first division grid as a*a and the size of the second division grid as b*b;

[0089] Step S202: Use the first division grid to perform grid division on the original background area. Denote the divided grid area as the background grid, and mark it in the order from left to right and from top to bottom. For the coordinates of any background grid, denote it as BW(x, y), where x represents the column number where the background grid is located, and y represents the row number where the background grid is located; Denote the original background area after completing the grid division as the grid background area;

[0090] Step S203: Use the second division grid to perform grid division on the original main area. Denote the divided grid area as the main grid, and mark it in the order from left to right and from top to bottom. For the coordinates of any main grid, denote it as RW(i, j), where i represents the column number where the main grid is located, and j represents the row number where the main grid is located; Denote the main background area after completing the grid division as the grid main area.

[0091] Step S3: Perform CT background desensitization processing on the grid background area based on the gray-scale distribution of the grid main body area to obtain a desensitized background area. Step S3 includes the following sub-steps:

[0092] Step S301: For any background grid in the grid background area, if it is adjacent to any main body grid, it is marked as a background adjacent grid, and the adjacent main body grid is marked as a main body adjacent grid; otherwise, it is marked as a background separated grid.

[0093] Step S304: Repeat to obtain all background separated grids, obtain the total number B0 of background separated grids, and randomly select B0*k1% of the background separated grids and denote them as the first background separated grids, and then randomly select B0*k2% of the background separated grids from the remaining background separated grids and denote them as the second background separated grids, where k1 and k2 are set proportionality coefficients.

[0094] Step S303: Randomly select B0*k1% of the main body grids and denote them as reference main body grids. For any reference main body grid, denoted as the first reference grid, arrange the gray-scale values corresponding to all pixel points in the first reference grid in ascending order and denote it as the reference gray-scale sequence.

[0095] Step S304: Obtain the average value of the smallest p1% of the gray-scale values in the reference gray-scale sequence and denote it as the first reference value, and obtain the average value of the largest p1% of the gray-scale values in the reference gray-scale sequence and denote it as the second reference value. Then obtain the average value of the reference gray-scale sequence and denote it as the third reference value. Randomly select one from the first reference value, the second reference value, and the third reference value and denote it as the reference gray-scale value of the corresponding reference main body grid. Repeat to obtain the reference gray-scale values of all reference main body grids and denote them as the reference gray-scale set.

[0096] Step S305: Perform first background desensitization processing on all the first background separated grids. Step S305 includes the following sub-steps:

[0097] Step S3051: Denote any one of the first background separated grids as the desensitized first grid, and obtain the pixel point at the geometric center of the desensitized first grid and denote it as the central pixel point.

[0098] [[ID=2I]]Step S3052: Draw a circle with the central pixel point as the center and 1 / 3*a as the radius and denote it as the first desensitized area.

[0099] Step S3053: Randomly select e1 pixel points from the first desensitized area and denote them as the first pixel points. Draw circles with the first pixel points as the centers and 1 / 6*a as the radii respectively and denote them as the second desensitized areas.

[0100] Step S3054: Randomly select a reference gray value from the reference gray value set, denoted as the desensitized gray value; change the gray values of the pixels in the first desensitized area and all the second desensitized areas to the desensitized gray value;

[0101] Step S306: Perform second background desensitization processing on all the second background separated grids, including: set the first random change value C1, C1 = -1, 0 or 1; for any one of the second background separated grids, denoted as the desensitized second grid, when C1 = -1, reduce the gray values of all the pixels in the second background separated grid by M1 respectively; when C1 = 1, increase the gray values of all the pixels in the second background separated grid by M1 respectively; when C1 = 0, no processing is performed;

[0102] Step S307: Mark the grid background area that has completed CT background desensitization processing as the desensitized background area.

[0103] Step S4: Perform CT main body desensitization processing on the grid main body area to obtain the desensitized main body area and obtain the desensitized CT image; Step S4 includes the following sub-steps:

[0104] Step S401: Mark the main body grids in the grid main body area that are not marked as main body adjacent grids as main body separated grids; obtain the total number B1 of the main body separated grids, and randomly select B1 * k1% of the main body separated grids and denote them as the first main body separated grids, and then randomly select B0 * k2% of the main body separated grids from the remaining background separated grids and denote them as the second main body separated grids;

[0105] Step S402: Set the second random change value C2, C2 = -1 or 1; perform first main body desensitization processing on all the first main body separated grids; Step S402 includes the following sub-steps:

[0106] Step S4021: Denote any one of the first main body separated grids as the main body first grid, and obtain the geometric center of the main body first grid, denoted as the first main body center;

[0107] Step S4022: Draw a circle with the first main body center as the center and k3 * 2 / b as the radius, denoted as the first main body circular area;

[0108] Step S4023: Randomly select k4% of the pixels from all the pixels included in the first main body circular area, denoted as the first desensitized pixels, and randomly assign a value to C2 as the marking value of all the first desensitized pixels; where k3 and k4 are set proportionality coefficients, 0 < {k3, k4} < 1;

[0109] Step S4024, set the grayscale size adjustment range [V1, V2]. For any first desensitized pixel, adjust the corresponding grayscale value according to the grayscale adjustment formula as follows: where AF is the grayscale value of the first desensitized pixel before adjustment; BF is the grayscale value of the first desensitized pixel after adjustment; repeat the adjustment for all first desensitized pixels;

[0110] Step S403, perform second subject desensitization processing on all second subject separated grids respectively; Step S403 includes the following sub-steps:

[0111] Step S4031, denote any second subject separated grid as the subject second grid, and obtain the geometric center of the subject second grid, denoted as the second subject center;

[0112] Step S4032, draw a circle with the second subject center as the center and k5*2 / b as the radius, denoted as the second subject circular area; then draw a circle with the second subject center as the center and k6*2 / b as the radius, denoted as the third subject circular area, where k5 and k6 are set proportionality coefficients, 1>k5>k6>0;

[0113] Step S4033, mark the annular area enclosed by the boundary of the second subject circular area and the boundary of the third subject circular area as the first annular area;

[0114] Step S4034, evenly divide the first annular area into k7 sector areas, denoted as the first sector areas, and randomly select k8 pixel points from each first sector area, denoted as desensitized edge pixels;

[0115] Step S4035, randomly select a desensitized edge pixel as the starting edge pixel, connect the desensitized edge pixel that is the closest to the starting edge pixel and is in the same first sector area as the starting edge pixel and has not been connected, and then use the connected desensitized edge pixel as the new starting edge pixel to continue connecting the next starting edge pixel;

[0116] Step S4036, if all desensitized edge pixels in the first sector area have been connected, then select the desensitized edge pixel that is the closest in the adjacent first sector area in the clockwise direction for connection until all desensitized edge pixels are connected, and connect the last connected desensitized edge pixel to the first starting edge pixel, and mark the formed closed area as the first subject desensitized area;

[0117] Step S4037, denote all pixel points in the first subject desensitized area as second desensitized pixels, and then randomly take a value for C2 as the marking value of all second desensitized pixels; then adjust the grayscale values of the second desensitized pixels respectively according to the grayscale adjustment formula;

[0118] Step S404: Mark the grid main body area that has completed the desensitization process of the CT main body as the desensitized main body area; and mark the corresponding original CT image as the desensitized CT image after completion.

[0119] Example 3, please refer to Figure 4 as shown in Figure 4 illustrates a schematic structural diagram of an electronic device, which may include: a processor, a communication interface, a memory, and a communication bus. Among them, the processor, the communication interface, and the memory complete mutual communication through the communication bus. The memory stores computer-readable instructions, and the processor can call the instructions in the memory. When the computer-readable instructions are executed by the processor, the steps in a desensitization processing method for privacy big data are run to implement the following functions: perform text erasure processing on the original CT image to obtain a first CT image, and divide it into an original main body area and an original background area; perform grid division on the original main body area and the original background area respectively to obtain a grid main body area and a grid background area; perform CT background desensitization processing on the grid background area based on the gray level distribution of the grid main body area to obtain a desensitized background area; perform CT main body desensitization processing on the grid main body area to obtain a desensitized main body area, and obtain a desensitized CT image.

[0120] In addition, when the logical instructions in the above-mentioned memory are implemented in the form of software functional units and sold or used as an independent product, they can be stored in a computer-readable storage medium. Based on such an understanding, the technical solution of the present application, in essence, or the part that contributes to the prior art, or a 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 may 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 application. The foregoing storage medium includes: various media such as a USB flash drive, a mobile hard disk, a read-only memory (ROM, Read-Only Memory), a random access memory (RAM, Random Access Memory), a magnetic disk, or an optical disc that can store program codes.

[0121] Embodiment 4. The present application further provides a computer-readable storage medium. The present application provides a storage medium on which a computer program is stored. When the computer program is executed by a processor, the steps in the above-mentioned desensitization processing method for privacy big data are run to implement the following functions: perform text erasure processing on the original CT image to obtain a first CT image, and divide it into an original main body area and an original background area; perform grid division on the original main body area and the original background area respectively to obtain a grid main body area and a grid background area; perform CT background desensitization processing on the grid background area based on the gray-scale distribution of the grid main body area to obtain a desensitized background area; perform CT main body desensitization processing on the grid main body area to obtain a desensitized main body area, and obtain a desensitized CT image.

[0122] Through the description of the above embodiments, the embodiments of the present invention can be provided as a method, a system or a computer program product. Based on such an understanding, the above technical solution, in essence, or the part that makes a contribution to the prior art can be embodied in the form of a software product. The computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disc, etc., and includes several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0123] In the embodiments provided in the present application, it should be understood that the disclosed system or method can be implemented in other ways. The above-described embodiments are merely illustrative. For example, the division of modules or units is only a logical function division. In actual implementation, there may be other division methods. For another example, multiple modules or units can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the displayed or discussed coupling or direct coupling or communication connection between each other can be through some communication interfaces. The indirect coupling or communication connection of systems, modules, and units can be in an electrical, mechanical, or other form.

[0124] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, and are not intended to limit them. Although the present application 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 the embodiments of the present application.

Claims

1. A method for desensitizing privacy big data, characterized in that, The method includes the following steps: Perform text erasure processing on the original CT image to obtain the first CT image, and divide it into the original main region and the original background region; Perform grid division on the original main region and the original background region respectively to obtain the grid main region and the grid background region; Based on the gray-scale distribution of the grid main region, perform CT background desensitization processing on the grid background region to obtain the desensitized background region; Perform CT main body desensitization processing on the grid main region to obtain the desensitized main region, and obtain the desensitized CT image.

2. The desensitization processing method for privacy big data according to claim 1, wherein Performing text erasure processing on the original CT image to obtain the first CT image, and dividing it into the original main region and the original background region includes the following sub-steps: For any text in the original CT image, denoted as the first text, obtain the horizontal and vertical coordinates of all pixel points contained in the first text in the original CT image, and arrange all the horizontal coordinates and all the vertical coordinates in ascending order, denoted as the horizontal coordinate sequence and the vertical coordinate sequence respectively; Obtain the minimum value and the maximum value in the horizontal coordinate sequence, denoted as XX and DX in sequence; obtain the minimum value and the maximum value in the vertical coordinate sequence, denoted as XY and DY in sequence; Connect the pixel points with coordinates [XX, XY], [DX, XY], [DX, DY], and [XX, DY] in the original CT image in sequence, and denote the enclosed closed area as the text erasure area. Change the gray-scale values of all pixel points in the text erasure area to K0; repeat the processing of all texts in the original CT image. After completion, obtain the first CT image; Divide the image area corresponding to the human tissue in the first CT image, denoted as the original main region, and denote the image area in the first CT image excluding the original main region and the text erasure area as the original background region.

3. The desensitization processing method for privacy big data according to claim 2, characterized in that, Performing grid division on the original main region and the original background region respectively to obtain the grid main region and the grid background region includes the following sub-steps: Set the size of the first division grid to a*a, and the size of the second division grid to b*b; Use the first division grid to perform grid division on the original background region, denote the divided grid area as the background grid, and mark it in the order from left to right and from top to bottom. For the coordinates of any background grid, denoted as BW(x, y), where x represents the column number where the background grid is located, and y represents the row number where the background grid is located; denote the original background region after completing the grid division as the grid background region; Use the second division grid to perform grid division on the original main region, denote the divided grid area as the main grid, and mark it in the order from left to right and from top to bottom. For the coordinates of any main grid, denoted as RW(i, j), where i represents the column number where the main grid is located, and j represents the row number where the main grid is located; denote the main background region after completing the grid division as the grid main region.

4. A desensitization processing method for privacy big data according to claim 3, characterized in that, Based on the gray-scale distribution of the grid main region, performing CT background desensitization processing on the grid background region to obtain the desensitized background region includes the following sub-steps: For any background grid in the grid background region, if it is adjacent to any main grid, it is marked as a background adjacent grid, and the adjacent main grid is marked as a main adjacent grid; otherwise, it is marked as a background separated grid. Repeat to obtain all background separated grids, obtain the total number B0 of background separated grids, and randomly select B0*k1% of the background separated grids and denote them as the first background separated grids. Then randomly select B0*k2% of the background separated grids from the remaining background separated grids and denote them as the second background separated grids, where k1 and k2 are set proportionality coefficients. Randomly select B0*k1% of the main grids and denote them as reference main grids. For any reference main grid, denoted as the first reference grid, arrange the gray values corresponding to all pixel points in the first reference grid in ascending order and denote it as the reference gray sequence. Obtain the average value of the smallest p1% of the gray values in the reference gray sequence and denote it as the first reference value, and obtain the average value of the largest p1% of the gray values in the reference gray sequence and denote it as the second reference value. Then obtain the average value of the reference gray sequence and denote it as the third reference value. Randomly select one from the first reference value, the second reference value, and the third reference value and denote it as the reference gray value of the corresponding reference main grid. Repeat to obtain the reference gray values of all reference main grids and denote them as the reference gray set.

5. The desensitization processing method for privacy big data according to claim 4, characterized in that, Performing CT background desensitization processing on the grid background region based on the gray distribution of the grid main region to obtain the desensitized background region further includes the following sub-steps: Perform first background desensitization processing on all the first background separated grids, including: Denote any one of the first background separated grids as the desensitized first grid, and obtain the pixel point at the geometric center of the desensitized first grid and denote it as the central pixel point. Draw a circle with the central pixel point as the center and 1 / 3*a as the radius and denote it as the first desensitized region. Randomly select e1 pixel points from the first desensitized region and denote them as the first pixel points. Draw circles with the first pixel points as the centers and 1 / 6*a as the radii respectively and denote them as the second desensitized regions. Randomly select a reference gray value from the reference gray set and denote it as the desensitized gray value. Change the gray values of the pixel points in the first desensitized region and all the second desensitized regions to the desensitized gray value. Perform second background desensitization processing on all the second background separated grids, including: Set the first random change value C1, where C1 = -1, 0, or 1. For any one of the second background separated grids, denoted as the desensitized second grid, when C1 = -1, reduce the gray values of all pixel points in the second background separated grid by M1 respectively; when C1 = 1, increase the gray values of all pixel points in the second background separated grid by M1 respectively; when C1 = 0, no processing is performed. Mark the grid background region that has completed CT background desensitization processing as the desensitized background region.

6. A desensitization processing method for private big data according to claim 5, characterized in that, Perform CT main desensitization processing on the grid main region to obtain the desensitized main region, and obtain the desensitized CT image, including the following sub-steps: Mark the main grids in the main grid area that are not marked as adjacent grids of the main body as separated grids of the main body; obtain the total number B1 of the separated grids of the main body, and randomly select B1*k1% of the separated grids of the main body and denote them as the first separated grids of the main body. Then, randomly select B0*k2% of the separated grids of the main body from the remaining background separated grids and denote them as the second separated grids of the main body; Set the second random variation value C2, where C2 = -1 or 1; perform the first main body desensitization processing on all the first separated grids of the main body, including: denote any first separated grid of the main body as the first main body grid, obtain the geometric center of the first main body grid and denote it as the first main body center; draw a circle with the first main body center as the center and k3*2 / b as the radius, and denote it as the first main body circular area; randomly select k4% of the pixel points from all the pixel points included in the first main body circular area and denote them as the first desensitized pixel points, and randomly assign a value to C2 as the marking value of all the first desensitized pixel points; where k3 and k4 are set proportionality coefficients, 0 < {k3, k4} < 1; Set the grayscale size adjustment range [V1, V2], and adjust the corresponding grayscale value for any first desensitized pixel point according to the grayscale adjustment formula as follows: Where AF is the grayscale value of the first desensitized pixel point before adjustment; BF is the grayscale value of the first desensitized pixel point after adjustment; Repeat the adjustment for all first desensitized pixel points.

7. A desensitization processing method for privacy big data according to claim 6, characterized in that Perform CT main body desensitization processing on the main grid area to obtain the desensitized main body area, and obtaining the desensitized CT image also includes the following sub-steps: Then perform the second main body desensitization processing on all the second separated grids of the main body, including: denote any second separated grid of the main body as the second main body grid, obtain the geometric center of the second main body grid and denote it as the second main body center; draw a circle with the second main body center as the center and k5*2 / b as the radius, and denote it as the second main body circular area; then draw a circle with the second main body center as the center and k6*2 / b as the radius, and denote it as the third main body circular area, where k5 and k6 are set proportionality coefficients, 1 > k5 > k6 > 0; mark the annular area enclosed by the boundary of the second main body circular area and the boundary of the third main body circular area as the first annular area; Evenly divide the first annular area into k7 sector areas and denote them as the first sector areas. Randomly select k8 pixel points from each first sector area and denote them as desensitized edge pixel points; randomly select a desensitized edge pixel point as the starting edge pixel point, connect the desensitized edge pixel point that is the closest to the starting edge pixel point and is in the same first sector area as the starting edge pixel point and has not been connected, and then use the connected desensitized edge pixel point as the new starting edge pixel point and continue to connect the next starting edge pixel point; if all the desensitized edge pixel points in the first sector area have been connected, then select the desensitized edge pixel point that is the closest in the adjacent first sector area in the clockwise direction and connect it until all the desensitized edge pixel points are connected, and connect the last connected desensitized edge pixel point to the first starting edge pixel point. Denote the formed closed area as the first main body desensitized area; Denote all the pixel points in the first main body desensitized area as the second desensitized pixel points, and randomly assign a value to C2 as the marking value of all the second desensitized pixel points; then adjust the gray values of the second desensitized pixel points respectively through the gray adjustment formula; Mark the grid main body area after the CT main body desensitization process as the desensitized main body area; and mark the corresponding original CT image as the desensitized CT image after completion.

8. A desensitization processing system for private big data, which is used to implement the desensitization processing method for private big data described in any one of claims 1-7, and is characterized in that, It includes an information division module, a grid division module, a background desensitization module, and a main body desensitization module; The information division module includes an information erasure unit and an area division unit; the information erasure unit is used to perform text erasure processing on the original CT image to obtain a first CT image, and the area division unit is used to divide the first CT image into an original main body area and an original background area; The grid division module is used to perform grid division on the original main body area and the original background area respectively to obtain a grid main body area and a grid background area; The background desensitization module performs CT background desensitization processing on the grid background area based on the gray level distribution of the grid main body area to obtain a desensitized background area; The main body desensitization module performs CT main body desensitization processing on the grid main body area to obtain a desensitized main body area and obtains a desensitized CT image.

9. An electronic device, characterized in that, It includes a processor and a memory, and the memory stores computer-readable instructions. When the computer-readable instructions are executed by the processor, the steps in the method according to any one of claims 1-7 are run.

10. A storage medium, on which a computer program is stored, characterized in that, When the computer program is executed by the processor, the steps in the method according to any one of claims 1-7 are run.

Images