An analysis and evaluation system for password security

Abstract

The present invention relates to the field of password security assessment, and in particular to an analysis and assessment system for password security, comprising: an image repository; a keyword matching unit for determining a keyword matching method for an evidence image according to an evidence image type; an image selection unit for determining an image selection method for a secret review text according to set conditions to select a target image; a risk determination unit for determining an image risk value determination method according to the evidence image type; a desensitization processing unit for determining a desensitization method for a target image according to analysis conditions; and an auxiliary selection unit for determining a supplementary selection method for an evidence image according to image relevance. The present invention can improve the integrity of assessment results while reducing the risk of information leakage.

Description

Technical Field

[0001] The present invention relates to the field of password security assessment, and in particular to an analysis and assessment system for password security. Background Art

[0002] When generating a confidential review report, the critical reviewer can place the confidential review-related evidence images in the confidential review report in the form of a hyperlink based on actual needs. However, at present, the access rights of the critical reviewers to the evidence images are mostly fully accessible, which makes the evidence images prone to the risk of leakage. Therefore, how to provide adaptive access rights to the evidence images according to the confidential review report on the basis of ensuring information security to improve the security of the confidential review process is a technical problem that needs to be urgently solved by those skilled in the art.

[0003] Chinese patent publication number CN117993844A discloses a method, device, medium, and electronic device for monitoring password applications. The method includes: using an intelligent solution recommendation algorithm to quickly match similar cases; business users compile password application solutions based on similar cases; business administrators evaluate the password application solutions through a pre-generated self-assessment model to determine an initial evaluation report; business administrators create a review process corresponding to the password application solution; review users fill in their review opinions on the password application solution and compile them into a review report; business administrators use OCR technology to parse and identify data in the review report, determine the core key data results of the review report, and determine whether the review report is passed based on the core key data results. It can be seen that the above technical solution has the following problems: the selection range of evidence images is limited, and thus the key image information related to the secret review text cannot be fully displayed, resulting in incomplete evaluation results, and the secret review report image may contain sensitive information, which poses a risk of information leakage. Summary of the Invention

[0004] To this end, the present invention provides an analysis and evaluation system for cryptographic security to overcome the problem in the prior art that the selection range of evidence images is limited, and thus the key image information related to the secret review text cannot be fully displayed, resulting in incomplete evaluation results, and sensitive information may exist in the secret review report image, posing a risk of information leakage.

[0005] To achieve the above objectives, the present invention provides a system for analyzing and evaluating password security, comprising:

[0006] Image repository for storing evidence images;

[0007] a keyword matching unit connected to the image repository and configured to determine a keyword matching method for the evidence image based on the type of the evidence image, wherein the keyword matching method includes determining matching keywords based on the number of times a pair is used or determining matching keywords based on a feature matching degree;

[0008] An image selection unit, connected to the keyword matching unit, is used to determine an image selection method for the review text according to set conditions to select a target image, wherein the image selection method is to select an evidence image according to a characteristic keyword or a keyword matching degree;

[0009] a risk determination unit, connected to the keyword matching unit and the image selection unit, respectively, for determining an image risk value determination method according to the type of evidence image, wherein the image risk value determination method is to determine the image risk value according to the area and number of sensitive areas or to determine the image risk value according to the number of sensitive keywords;

[0010] a desensitization processing unit, which is connected to the image selection unit and the risk determination unit respectively, and is used to determine a desensitization method for the target image according to the analysis conditions, wherein the desensitization method is to select a reference number of sensitive features in each feature combination for fuzzy processing or to perform fuzzy processing on all sensitive features;

[0011] An auxiliary selection unit is respectively connected to the image storage library, the image selection unit and the desensitization processing unit, and is used to determine a supplementary selection method for the evidence image according to the image relevance. The supplementary selection method is to select supplementary images in descending order of feature association coefficients or keyword matching coefficients.

[0012] Further, the keyword matching unit is responsive to the evidence image type to determine a keyword matching method for the evidence image;

[0013] The keyword matching unit responds that the evidence image type is a text evidence image, and determines the keyword matching method as determining the matching keyword based on the number of times the combination is used;

[0014] The keyword matching unit responds that the evidence image type is a scene evidence image, and determines the keyword matching method as determining the matching keyword based on the feature matching degree;

[0015] The evidence image types include text evidence images and scene evidence images.

[0016] Furthermore, the keyword matching unit determines matching keywords based on the number of times the combination is used, including:

[0017] For a text evidence image, the text evidence image is recorded as a target text evidence image, and other text evidence images other than the target text evidence image are recorded as reference text evidence images, and each similar image corresponding to the target text evidence image is detected, and matching analysis is performed on each similar image;

[0018] When performing matching analysis on a single similar image, the similar image is recorded as the reference similar image, the same keywords in the target text evidence image and the reference similar image are recorded as reference same keywords, the other keywords in the target text evidence image other than the reference keywords are recorded as keywords to be analyzed, and the keywords in the uploaded data that are used in combination with the single keyword to be analyzed more than the preset number of times are recorded as collocation keywords;

[0019] Each keyword in the target text evidence image and the collocation keyword corresponding to each keyword to be analyzed are recorded as the matching keyword of the target text evidence image;

[0020] The similar image is a reference text evidence image in which the number of identical keywords with the target text evidence image is greater than a preset number of identical keywords and the distribution similarity of identical keywords is greater than a preset distribution similarity of identical keywords.

[0021] Furthermore, the image selection unit is responsive to the set conditions to determine a method for selecting images of the secret review text;

[0022] The setting condition of the image selection unit response is that the proportion of the influencing keywords is greater than or equal to the preset proportion of the influencing keywords or the influencing factor is greater than or equal to the preset influencing factor, and the image selection method is to select the evidence image according to the characteristic keywords;

[0023] The setting condition of the image selection unit response is that the proportion of the influencing keywords is less than the preset proportion of the influencing keywords and the influencing factor is less than the preset influencing factor, and the image selection method is to select the evidence image according to the keyword matching degree;

[0024] The characteristic keywords are keywords in the influence keyword group whose sub-influence factors are greater than the preset sub-influence factors.

[0025] Furthermore, the image selection unit determines the sub-influence factor according to the disorder coefficient, wherein,

[0026] If the disorder coefficient is greater than or equal to the preset disorder coefficient, the sub-influence factor is determined according to the influence mean;

[0027] If the disorder coefficient is less than the preset disorder coefficient, the sub-influence factor is determined according to the paragraph length and the number of associations.

[0028] Furthermore, the risk determination unit responds to the evidence image type to determine the image risk value determination method;

[0029] The risk determination unit corresponds to the scene evidence image and determines the image risk value according to the area and number of sensitive areas;

[0030] The risk determination unit corresponds to the text evidence image and determines the image risk value based on the number of sensitive keywords.

[0031] Furthermore, the desensitization processing unit responds to the analysis conditions to determine a desensitization method for the target image;

[0032] The analysis condition for the response of the desensitization processing unit is that the proportion of risk images is greater than or equal to the preset proportion of risk images or the mean value of the sensitive feature distribution is greater than or equal to the preset mean value of the sensitive feature distribution. The desensitization method is to select a reference number of sensitive features in each feature combination for fuzzy processing;

[0033] The analysis conditions for the response of the desensitization processing unit are that the proportion of risk images is less than the preset proportion of risk images and the mean value of the sensitive feature distribution is less than the preset mean value of the sensitive feature distribution, and the desensitization method is determined to be fuzzy processing for all sensitive features.

[0034] Furthermore, the desensitization method of the desensitization processing unit response is to select a reference number of sensitive features in each feature combination for fuzzy processing, wherein,

[0035] When performing a combined analysis on a single sensitive feature in a single evidence image, the sensitive feature is recorded as a target sensitive feature, and other sensitive features in the evidence image other than the target sensitive feature are recorded as reference sensitive features. The set of reference sensitive features and the target sensitive features whose feature correlation with the target sensitive feature is greater than the preset feature correlation is recorded as a feature combination;

[0036] Continue to perform combined analysis on sensitive features that are not recorded in the feature combination until all sensitive features are recorded in the feature combination. Randomly select a reference number of sensitive features in a single feature combination for fuzzy processing. The reference number is positively correlated with the number of sensitive features corresponding to the single feature combination.

[0037] The sensitive features include sensitive areas and sensitive keywords.

[0038] Furthermore, the desensitization processing unit determines a feature correlation degree based on the sensitive features;

[0039] If the sensitive feature is a sensitive keyword, the feature relevance is determined based on the collocation coefficient and the interval reference value;

[0040] If the sensitive feature is a sensitive area, the feature association degree is determined based on the device association degree.

[0041] Furthermore, the auxiliary selection unit determines a supplementary selection method for the evidence image based on the image correlation;

[0042] If the image correlation is greater than or equal to the preset image correlation, the supplementary selection method is to select supplementary images in descending order of feature correlation coefficients;

[0043] If the image relevance is less than the preset image relevance, the supplementary selection method is to select supplementary images in descending order of keyword matching coefficients.

[0044] Compared with the prior art, the beneficial effect of the present invention lies in that, in the technical solution of the present invention, different keyword matching methods for evidence images are adaptively selected according to the type of evidence image, so that the selection of keyword matching method is more in line with the actual state of generating matching keywords, and can more accurately capture the key information in the evidence image, avoiding the problem of poor correlation between the subsequent recommended evidence images and the secret review text due to unreasonable matching keyword settings, ensuring that the recommended evidence images are closely related to the secret review text, helping to improve the overall quality of the secret review report, and can comprehensively display the key image information related to the secret review text, thereby improving the persuasiveness and credibility of the evaluation results.

[0045] Furthermore, the present invention determines the image selection method of the secret review text according to the set conditions, and effectively reflects the influence degree of the influencing keywords in the text evidence image and the actual characteristics of the text evidence image by setting the conditions, and then adaptively selects different image selection methods according to the set conditions. It can automatically adjust the selection strategy according to the specific characteristics of the text evidence image, so that the selected evidence image has a higher degree of correlation with the secret review text, thereby improving the flexibility and pertinence of the evidence image selection.

[0046] Furthermore, the present invention effectively reflects the distribution status of the same influencing keywords through the confusion coefficient, and then determines the sub-influence factor based on the confusion coefficient, so that the determination of the sub-influence coefficient is more in line with the actual distribution characteristics, thereby improving the accuracy of the determination of the sub-influence factor, which is beneficial to the subsequent selection of evidence images based on feature keywords, thereby improving the accuracy of the selection of evidence images.

[0047] Furthermore, the present invention determines the desensitization method of the target image based on the analysis conditions, effectively reflects the proportion of risk images and the distribution of sensitive features through the analysis conditions, and then adaptively selects different desensitization methods according to the analysis conditions, so that the selection of desensitization method is more in line with the actual application scenario. By effectively desensitizing sensitive information, the risk of data leakage can be reduced and the security of confidentiality review work can be improved.

[0048] Furthermore, the present invention determines the supplementary selection method of the evidence image based on the image correlation, and effectively reflects the correlation degree of the selected target image through the image correlation, and then adaptively selects different supplementary selection methods according to the image correlation, so that the supplementary selection method can select evidence images with a higher degree of correlation with the target image, thereby enhancing the integrity of the evidence image in the secret review report. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] Figure 1A unit connection diagram of the cryptographic security analysis and evaluation system of the present invention;

[0050] Figure 2 This is a flow chart of the present invention for determining a keyword matching method for an evidence image according to the type of the evidence image;

[0051] Figure 3 This is a flow chart of a method for selecting images of a secret review text according to set conditions in the present invention;

[0052] Figure 4 This is a flow chart of the present invention for determining a desensitization method for a target image based on analysis conditions. DETAILED DESCRIPTION

[0053] In order to make the objects and advantages of the present invention more clearly understood, the present invention is further described below in conjunction with embodiments; it should be understood that the specific embodiments described herein are merely used to explain the present invention and are not intended to limit the present invention.

[0054] The preferred embodiments of the present invention are described below with reference to the accompanying drawings. It should be understood by those skilled in the art that these embodiments are only used to explain the technical principles of the present invention and are not intended to limit the scope of protection of the present invention.

[0055] It should be noted that, in the description of the present invention, terms such as "up", "down", "left", "right", "inside", and "outside" indicating directions or positional relationships are based on the directions or positional relationships shown in the accompanying drawings. This is only for the convenience of description and does not indicate or imply that the device or element must have a specific orientation, be constructed and operated in a specific orientation. Therefore, it cannot be understood as a limitation on the present invention.

[0056] Furthermore, it should be noted that, in the description of the present invention, unless otherwise expressly specified or limited, the terms "mounted," "connected," and "connected" should be understood in a broad sense. For example, they may refer to fixed connections, detachable connections, or integral connections; mechanical connections or electrical connections; direct connections or indirect connections through an intermediate medium; and internal communication between two components. Those skilled in the art will understand the specific meanings of the above terms in the present invention based on specific circumstances.

[0057] See also Figures 1 to 4 As shown, the present invention provides an analysis and evaluation system for password security, comprising:

[0058] Image repository for storing evidence images;

[0059] a keyword matching unit connected to the image repository and configured to determine a keyword matching method for the evidence image based on the type of the evidence image, wherein the keyword matching method includes determining matching keywords based on the number of times a pair is used or determining matching keywords based on a feature matching degree;

[0060] An image selection unit, connected to the keyword matching unit, is used to determine an image selection method for the review text according to set conditions to select a target image, wherein the image selection method is to select an evidence image according to a characteristic keyword or a keyword matching degree;

[0061] a risk determination unit, connected to the keyword matching unit and the image selection unit, respectively, for determining an image risk value determination method according to the type of evidence image, wherein the image risk value determination method is to determine the image risk value according to the area and number of sensitive areas or to determine the image risk value according to the number of sensitive keywords;

[0062] a desensitization processing unit, which is connected to the image selection unit and the risk determination unit respectively, and is used to determine a desensitization method for the target image according to the analysis conditions, wherein the desensitization method is to select a reference number of sensitive features in each feature combination for fuzzy processing or to perform fuzzy processing on all sensitive features;

[0063] An auxiliary selection unit is respectively connected to the image storage library, the image selection unit and the desensitization processing unit, and is used to determine a supplementary selection method for the evidence image according to the image relevance. The supplementary selection method is to select supplementary images in descending order of feature association coefficients or keyword matching coefficients.

[0064] Among them, the application scenario of the present invention is the recommendation of evidence images in the secret review text generated during the password security assessment process. The secret review text is the text after the password security is evaluated. In the present invention, several historical records are correspondingly set. Any historical record records the feature matching degree, the number of the same keywords, the similarity of the distribution of the same keywords, the influencing factor, the keyword matching degree, the sub-influence factor and the confusion coefficient in the historical process of at least one evidence image recommendation, and each historical record corresponds to a qualified mark. The qualified mark records whether the stability detection process meets the user requirements. The qualified mark can be recorded manually. It can be understood that the user can determine whether the stability detection process meets the requirements based on self-set indicators. The self-set indicators can be but not limited to the error recommendation rate, which will not be elaborated here. The error recommendation rate is the number of times the evidence image is incorrectly recommended.

[0065] Specifically, the keyword matching unit responds to the evidence image type to determine a keyword matching method for the evidence image;

[0066] The keyword matching unit responds that the evidence image type is a text evidence image, and determines the keyword matching method as determining the matching keyword based on the number of times the combination is used;

[0067] The keyword matching unit responds that the evidence image type is a scene evidence image, and determines the keyword matching method as determining the matching keyword based on the feature matching degree;

[0068] The evidence image types include text evidence images and scene evidence images.

[0069] When determining the matching keywords based on the feature matching degree, the keywords in the uploaded data with a feature matching degree greater than the preset feature matching degree are selected as the matching keywords of the evidence image;

[0070] Evidence image types include text evidence images and scene evidence images. Text evidence images are images with text appearing, while scene evidence images are images without text appearing.

[0071] The method for confirming the number of times of combination use is as follows: for a single keyword to be analyzed, the keyword to be analyzed is recorded as the target keyword to be analyzed, the position where the target keyword to be analyzed appears in the uploaded data is detected, and the adjacent keywords corresponding to the target keywords to be analyzed at different positions are detected; for a single adjacent keyword, the adjacent keyword is recorded as the target keyword, and the number of times the target keyword appears in the adjacent keywords corresponding to each target keyword to be analyzed appearing in the uploaded data is recorded as the number of times of combination use corresponding to the target keyword; for a single keyword, the adjacent keyword corresponding to the keyword is a keyword whose character interval with the keyword is less than a first preset character interval; for two arbitrary keywords, the character interval is the number of characters between the two keywords, and the uploaded data is the text uploaded by the user to the password security detection and evaluation system;

[0072] The value of the first preset character interval can be determined by the user according to the actual application scenario. The higher the user's requirement for the relevance between the collocation keyword and the keyword to be analyzed, the smaller the value of the first preset character interval. A value of the first preset character interval is provided, and the first preset character interval is 4;

[0073] The method for confirming the feature matching degree is as follows: for a single scene evidence image, the name corresponding to each network device in the scene evidence image is identified and recorded as a name keyword; for a single keyword in the uploaded data, the keyword is recorded as a target keyword, and the feature matching degree corresponding to the target keyword = 1 / the average value of the minimum character interval between the target keyword and each name keyword in the uploaded data; the method for confirming the minimum character interval between the target keyword and a single name keyword in the uploaded data is to record the name keyword as the target name keyword, record the target name keywords at different positions in the uploaded data as reference keywords, detect the character interval between the target keyword and each reference keyword, and record the minimum value of each character interval as the minimum character interval;

[0074] In the present invention, network devices include servers, network devices, and storage devices. Network devices include but are not limited to switches, routers, and firewalls. Servers include but are not limited to hard disk arrays, tape libraries, and NAS. Users can identify the names corresponding to network devices through machine vision and deep learning networks. This is content that is easy for those skilled in the art to understand and will not be described in detail.

[0075] The value of the preset feature matching degree can be determined by the user according to the actual application scenario. The greater the degree of matching between the user's matching keywords and the scene evidence image, the greater the value of the preset feature matching degree. A value of the preset feature matching degree is provided, and the historical records of keyword matching based on the feature matching degree are detected. The average value of the feature matching degrees corresponding to the historical records that can meet the user's needs is recorded as the preset feature matching degree.

[0076] Specifically, the keyword matching unit determines the matching keyword according to the number of times the combination is used, including:

[0077] For a text evidence image, the text evidence image is recorded as a target text evidence image, and other text evidence images other than the target text evidence image are recorded as reference text evidence images, and each similar image corresponding to the target text evidence image is detected, and matching analysis is performed on each similar image;

[0078] When performing matching analysis on a single similar image, the similar image is recorded as the reference similar image, the same keywords in the target text evidence image and the reference similar image are recorded as reference same keywords, the other keywords in the target text evidence image other than the reference keywords are recorded as keywords to be analyzed, and the keywords in the uploaded data that are used in combination with the single keyword to be analyzed more than the preset number of times are recorded as collocation keywords;

[0079] Each keyword in the target text evidence image and the collocation keyword corresponding to each keyword to be analyzed are recorded as the matching keyword of the target text evidence image;

[0080] The similar image is a reference text evidence image in which the number of identical keywords with the target text evidence image is greater than a preset number of identical keywords and the distribution similarity of identical keywords is greater than a preset distribution similarity of identical keywords.

[0081] The method for confirming the number of identical keywords and the similarity of identical keyword distribution is as follows: for any two text evidence images, the number of identical keywords is the total number of identical keywords in the two text evidence images, the larger value of the sub-coefficients corresponding to the two text evidence images is recorded as a1, and the smaller value is recorded as a2, and the similarity of identical keyword distribution = 1 / (a1-a2), the sub-coefficient corresponding to a single text evidence image is the average value of the reference distances corresponding to the identical keywords in the text evidence image, the reference distance corresponding to the identical keyword is the shortest distance from the center position of the identical keyword to the reference point, the reference point is the center of the circumscribed circle of the text evidence image, and the center position of a single keyword is the center of the circumscribed circle of the smallest rectangle that can contain the keyword;

[0082] The values of the preset number of times a combination is used, the preset number of identical keywords, and the preset similarity of the distribution of identical keywords can be determined by the user according to the actual application scenario. The higher the user's demand for the relevance between the combination keywords and the keywords to be analyzed, the larger the value of the preset number of times a combination is used. A value of a preset number of times a combination is used is provided, and the preset number of times a combination is used is 5 times. The higher the user's demand for the similarity of similar images of the same type, the larger the values of the preset number of identical keywords and the preset similarity of the distribution of identical keywords. A value of a preset number of identical keywords and the preset similarity of the distribution of identical keywords is provided. The historical records of similar images of the same type are determined based on the number of identical keywords and the similarity of the distribution of identical keywords. The average value of the number of identical keywords corresponding to the historical records that can meet the user's needs is recorded as the preset number of identical keywords, and the average value of the similarity of the distribution of identical keywords corresponding to the historical records that can meet the user's needs is recorded as the preset similarity of the distribution of identical keywords.

[0083] Specifically, the image selection unit responds to the set conditions to determine the image selection method of the secret review text;

[0084] The setting condition of the image selection unit response is that the proportion of the influencing keywords is greater than or equal to the preset proportion of the influencing keywords or the influencing factor is greater than or equal to the preset influencing factor, and the image selection method is to select the evidence image according to the characteristic keywords;

[0085] The setting condition of the image selection unit response is that the proportion of the influencing keywords is less than the preset proportion of the influencing keywords and the influencing factor is less than the preset influencing factor, and the image selection method is to select the evidence image according to the keyword matching degree;

[0086] The characteristic keywords are keywords in the influence keyword group whose sub-influence factors are greater than the preset sub-influence factors.

[0087] The setting conditions include a first setting condition and a second setting condition. The first setting condition is that the proportion of influencing keywords is less than the preset proportion of influencing keywords and the influencing factor is less than the preset influencing factor. The second setting condition is that the proportion of influencing keywords is less than the preset proportion of influencing keywords and the influencing factor is less than the preset influencing factor.

[0088] When selecting evidence images based on characteristic keywords, the evidence images whose keyword matching degree with the secret review text is greater than the preset keyword matching degree are used as the target images corresponding to the secret review text;

[0089] When selecting evidence images based on keyword matching, the characteristic keywords in the review text are identified, and the evidence images containing the characteristic keywords are used as the target images corresponding to the review text;

[0090] The proportion of influential keywords = the number of influential keyword groups / (the number of keyword groups + the number of keywords with a count of 1 in the review text). Influential keywords are keywords with a count of more than a first preset count in the review text. The set of keywords with a count of more than 1 in the review text is recorded as a keyword group. If a keyword in a keyword group is an influential keyword, then the keyword group is considered an influential keyword group. The influence factor is the average of the sub-influence factors corresponding to each influential keyword group.

[0091] The values of the first preset number of occurrences, the preset proportion of influencing keywords and the preset influence factor can be determined by the user according to the actual application scenario. The smaller the value of the first preset number of occurrences, the greater the user's demand for determining the keyword as an influencing keyword. A value of the first preset number of occurrences is provided, and the first preset number of occurrences is 5 times. The smaller the values of the preset proportion of influencing keywords and the preset influence factor are, the greater the user's demand for selecting evidence images according to feature keywords. A value of the preset proportion of influencing keywords and the preset influence factor is provided, and the preset proportion of influencing keywords is 50%. The historical records of selecting evidence images according to feature keywords are detected, and the average value of the influence factors corresponding to the historical records that can meet the user's needs is recorded as the preset influence factor;

[0092] The method for confirming the keyword matching degree is as follows: for an evidence image, record the evidence image as the target evidence image, and detect whether the matching keyword corresponding to the target evidence image appears in the secret review text. If the matching keyword appears in the secret review text, it is recorded as the target matching keyword. The keyword matching degree = the number of target matching keywords in the matching keywords corresponding to the target evidence image / the total number of matching keywords corresponding to the target evidence image;

[0093] The values of the preset keyword matching degree and the preset sub-influence factor can be determined by the user according to the actual application scenario. The greater the user's demand for improving the matching degree between the evidence image and the secret review text, the greater the values of the preset keyword matching degree and the preset sub-influence factor. A value of the preset keyword matching degree is provided, and the historical records of selecting evidence images based on the keyword matching degree are detected. The average value of the keyword matching degree corresponding to the historical records that can meet the user's needs is recorded as the preset keyword matching degree. A value of the preset sub-influence factor is provided, and the historical records of selecting evidence images based on the feature keywords are detected. The average value of the sub-influence factors corresponding to the historical records that can meet the user's needs is recorded as the preset sub-influence factor.

[0094] Specifically, the image selection unit determines the sub-influence factor according to the disorder coefficient, wherein,

[0095] If the disorder coefficient is greater than or equal to the preset disorder coefficient, the sub-influence factor is determined according to the influence mean;

[0096] If the disorder coefficient is less than the preset disorder coefficient, the sub-influence factor is determined according to the paragraph length and the number of associations.

[0097] Among them, if the disorder coefficient is greater than or equal to the preset disorder coefficient, the sub-influence factor is positively correlated with the influence mean; if the disorder coefficient is less than the preset disorder coefficient, the sub-influence factor = paragraph length - number of associations;

[0098] The confusion coefficient is confirmed by, for a single influencing keyword in a single influencing keyword group, recording the influencing keyword as the target influencing keyword, recording the other influencing keywords in the influencing keyword group except the target influencing keyword as the reference influencing keywords, and recording the average value of the shortest distance from the center position of the target influencing keyword to the center position of each reference influencing keyword as the confusion coefficient corresponding to the influencing keyword group;

[0099] The impact mean is the average value of the impact threshold corresponding to each influencing keyword in a single influencing keyword group, the impact threshold is the total number of associated keywords corresponding to a single influencing keyword, and the method for confirming the associated keywords corresponding to a single influencing keyword is as follows: for a single influencing keyword, the influencing keyword is recorded as the target keyword, and the influencing keyword whose character interval with the target keyword is less than the second preset character interval is recorded as the associated keyword corresponding to the target keyword; the paragraph length is the number of characters between the first influencing keyword in a single influencing keyword group and the last influencing keyword in the keyword group in the order of reading the secret review text from front to back, and the associated number is the number of times the adjacent keyword with the largest number of occurrences appears among the adjacent keywords corresponding to each influencing keyword in the single influencing keyword group;

[0100] The values of the preset confusion coefficient and the second preset character interval can be determined by the user according to the actual application scenario. The smaller the value of the preset confusion coefficient, the greater the user's demand for determining the sub-influence factor based on the influence mean. A value of the preset confusion coefficient is provided, and the historical records of determining the sub-influence factor based on the influence mean are detected. The average value of the confusion coefficient corresponding to the historical records that can meet the user's needs is recorded as the preset confusion coefficient. The greater the user's demand for the effectiveness of providing associated keywords on the influencing keywords, the smaller the value of the second preset character interval is. A value of the second preset character interval is provided, and the value of the second preset character interval is 10.

[0101] Specifically, the risk determination unit responds to the evidence image type to determine the image risk value determination method;

[0102] The risk determination unit corresponds to the scene evidence image and determines the image risk value according to the area and number of sensitive areas;

[0103] The risk determination unit corresponds to the text evidence image and determines the image risk value based on the number of sensitive keywords.

[0104] For a single scene evidence image, the image risk value = sensitive area + number of sensitive areas;

[0105] For text evidence images, the image risk value is positively correlated with the number of sensitive keywords;

[0106] The area of the sensitive region is the sum of the areas of all sensitive regions in a single scene evidence image. Users can determine the area of each sensitive region through OpenCV. The details are not detailed here. The number of sensitive regions is the total number of sensitive regions in a single scene evidence image. The sensitive region is the area where each network device is located in the scene evidence image.

[0107] The number of sensitive keywords is the total number of sensitive keywords in a single text evidence image. Sensitive keywords are keywords that appear in the uploaded data more than the second preset number of occurrences. The value of the second preset number of occurrences can be determined by the user based on the actual application scenario. The greater the user's demand for sensitivity judgment of sensitive keywords, the smaller the value of the second preset number of occurrences. A value of the second preset number of occurrences is provided, and the second preset number of occurrences is 15 times.

[0108] Specifically, the desensitization processing unit responds to the analysis conditions to determine a desensitization method for the target image;

[0109] The analysis condition for the response of the desensitization processing unit is that the proportion of risk images is greater than or equal to the preset proportion of risk images or the mean value of the sensitive feature distribution is greater than or equal to the preset mean value of the sensitive feature distribution. The desensitization method is to select a reference number of sensitive features in each feature combination for fuzzy processing;

[0110] The analysis conditions for the response of the desensitization processing unit are that the proportion of risk images is less than the preset proportion of risk images and the mean value of the sensitive feature distribution is less than the preset mean value of the sensitive feature distribution, and the desensitization method is determined to be fuzzy processing for all sensitive features.

[0111] The analysis conditions include a first analysis condition and a second analysis condition. The first analysis condition is that the proportion of risk images is greater than or equal to the preset proportion of risk images or the mean of the sensitive feature distribution is greater than or equal to the preset mean of the sensitive feature distribution. The second analysis condition is that the proportion of risk images is less than the preset proportion of risk images and the mean of the sensitive feature distribution is less than the preset mean of the sensitive feature distribution.

[0112] The proportion of risk images = the total number of risk evidence images / the number of all evidence images. Risk images are evidence images whose image risk value is greater than the preset image risk value.

[0113] The mean value of the sensitive feature distribution is the average value of the sensitive feature distribution coefficients corresponding to each evidence image. The sensitive feature distribution coefficient is confirmed as follows: for a single evidence image, the evidence image is recorded as the target evidence image, the sensitive feature distribution coefficient corresponding to the target evidence image is the average value of the sensitive distances corresponding to each sensitive feature in the target evidence image, for a single sensitive feature, the sensitive feature is recorded as the target sensitive feature, other sensitive features other than the target sensitive feature in the target evidence image are recorded as reference features, and the minimum value of the shortest distances from the center position of the target sensitive feature to the center position of each reference sensitive feature is recorded as the sensitive distance corresponding to the target sensitive feature; if the sensitive feature is a sensitive area, the center position of the sensitive feature is the center point of the circumscribed circle of the single sensitive area; if the sensitive feature is a keyword, the center position of the sensitive feature is the center point of the circumscribed circle of the minimum rectangle that can contain the single keyword;

[0114] The values of the preset risk image ratio, the preset sensitive feature distribution coefficient, and the preset image risk value can be determined by the user according to the actual application scenario. The larger the values of the preset risk image ratio and the preset sensitive feature distribution coefficient, the greater the user's demand for blurring all sensitive features. A value of the preset risk image ratio and the preset sensitive feature distribution coefficient is provided. The preset risk image ratio is 50%. The historical records of the reference number of sensitive features in each feature combination for blurring are detected and selected, and the average value of the sensitive feature distribution coefficients corresponding to the historical records that can meet the user's needs is recorded as the preset sensitive feature distribution coefficient;

[0115] When performing fuzzy processing on sensitive features, a fuzzification algorithm is used to make the sensitive features blurred and difficult to identify. The fuzzification algorithm includes but is not limited to fuzzy logic and fuzzy set theory. Users can choose according to actual needs, and the details are not described here.

[0116] Specifically, the desensitization method of the desensitization processing unit response is to select a reference number of sensitive features in each feature combination for fuzzy processing, wherein,

[0117] When performing a combined analysis on a single sensitive feature in a single evidence image, the sensitive feature is recorded as a target sensitive feature, and other sensitive features in the evidence image other than the target sensitive feature are recorded as reference sensitive features. The set of reference sensitive features and the target sensitive features whose feature correlation with the target sensitive feature is greater than the preset feature correlation is recorded as a feature combination;

[0118] Continue to perform combined analysis on sensitive features that are not recorded in the feature combination until all sensitive features are recorded in the feature combination. Randomly select a reference number of sensitive features in a single feature combination for fuzzy processing. The reference number is positively correlated with the number of sensitive features corresponding to the single feature combination.

[0119] The sensitive features include sensitive areas and sensitive keywords.

[0120] Among them, the value of the preset feature correlation degree can be determined by the user according to the actual application scenario. The greater the user's demand for the correlation degree of each sensitive feature in the feature combination, the larger the value of the preset feature correlation degree is. A value of the preset feature correlation degree is provided, and the historical records of a reference number of sensitive features in each feature combination are detected and selected for fuzzy processing. The average value of the feature correlation degrees corresponding to the historical records that can meet the user's needs is recorded as the preset feature correlation degree.

[0121] Specifically, the desensitization processing unit determines the feature correlation degree according to the sensitive features;

[0122] If the sensitive feature is a sensitive keyword, the feature relevance is determined based on the collocation coefficient and the interval reference value;

[0123] If the sensitive feature is a sensitive area, the feature association degree is determined based on the device association degree.

[0124] Among them, for sensitive keywords, feature correlation = collocation coefficient - interval reference value; for sensitive areas, feature correlation is positively correlated with device correlation; the method of confirming the collocation coefficient is to record the two sensitive keywords in a single text evidence image as the first keyword and the second keyword in the order of reading from first to last, detect the position where the first keyword appears in the uploaded data, and detect the adjacent keywords corresponding to the target sensitive keywords at different positions, and record the number of times the second keyword appears in each adjacent keyword as the collocation coefficient corresponding to the two sensitive keywords; the interval reference value is the number of characters between the first keyword and the second keyword in a single text evidence image; device correlation = 1 / the shortest distance between the center positions corresponding to the two sensitive areas.

[0125] Specifically, the auxiliary selection unit determines a supplementary selection method for the evidence image based on the image relevance;

[0126] If the image correlation is greater than or equal to the preset image correlation, the supplementary selection method is to select supplementary images in descending order of feature correlation coefficients;

[0127] If the image relevance is less than the preset image relevance, the supplementary selection method is to select supplementary images in descending order of keyword matching coefficients.

[0128] Among them, the evidence image that is not selected as the target image is recorded as the reference image. When selecting the supplementary image in the order of feature correlation coefficient from large to small and the supplementary image in the order of keyword matching coefficient from large to small, the reference image is selected as the supplementary image. The number of supplementary images selected can be determined by the user according to actual needs and is not limited.

[0129] Image correlation = feature correlation - Smin / Smax, where Smin and Smax are the smaller and larger values of the number of text evidence images and scene evidence images in the selected target image, respectively. The maximum value of the feature correlation in the selected target image is recorded as m1, and the minimum value is recorded as m2. Feature correlation = 1-(m1-m2) / m1. The feature correlation coefficient is determined as follows: for a single reference image, the feature correlation coefficient corresponding to the reference image = 1 / |feature correlation corresponding to the reference image - average value of the feature correlations corresponding to each target image|.

[0130] The keyword matching coefficient is determined by recording the matching keywords corresponding to the target image as analysis keywords. For a single reference image, the keyword matching coefficient corresponding to the reference image = the number of matching keywords corresponding to the reference image that are identical to the analysis keywords / the total number of matching keywords corresponding to the reference image.

[0131] The value of the preset image correlation can be determined by the user according to the actual application scenario. The smaller the value of the preset image correlation, the greater the user's demand for selecting supplementary images in descending order of feature correlation coefficients. A value of the preset image correlation is provided, and the historical records of selecting supplementary images in descending order of feature correlation coefficients are detected. The average value of the image correlation corresponding to the historical records that can meet the user's needs is recorded as the preset image correlation.

[0132] Thus far, the technical solutions of the present invention have been described in conjunction with the preferred embodiments shown in the accompanying drawings. However, it will be readily understood by those skilled in the art that the scope of protection of the present invention is obviously not limited to these specific embodiments. Without departing from the principles of the present invention, those skilled in the art may make equivalent changes or substitutions to the relevant technical features, and the technical solutions after such changes or substitutions will fall within the scope of protection of the present invention.

[0133] The foregoing description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Those skilled in the art will readily appreciate that the present invention is susceptible to various modifications and variations. Any modifications, equivalent substitutions, or improvements made within the spirit and principles of the present invention are intended to be within the scope of protection of the present invention.

Claims

1. A cryptographic security analysis and evaluation system, characterized in that: include: Image repository for storing evidence images; a keyword matching unit connected to the image repository and configured to determine a keyword matching method for the evidence image based on the type of the evidence image, wherein the keyword matching method includes determining matching keywords based on the number of times a pair is used or determining matching keywords based on a feature matching degree; An image selection unit, connected to the keyword matching unit, is used to determine an image selection method for the review text according to set conditions to select a target image, wherein the image selection method is to select an evidence image according to a characteristic keyword or a keyword matching degree; a risk determination unit, connected to the keyword matching unit and the image selection unit, respectively, for determining an image risk value determination method according to the type of evidence image, wherein the image risk value determination method is to determine the image risk value according to the area and number of sensitive areas or to determine the image risk value according to the number of sensitive keywords; a desensitization processing unit, which is connected to the image selection unit and the risk determination unit respectively, and is used to determine a desensitization method for the target image according to the analysis conditions, wherein the desensitization method is to select a reference number of sensitive features in each feature combination for fuzzy processing or to perform fuzzy processing on all sensitive features; an auxiliary selection unit, connected to the image storage, the image selection unit, and the desensitization processing unit, respectively, for determining a supplementary selection method for the evidence image based on image relevance, wherein the supplementary selection method selects supplementary images in descending order of feature correlation coefficients or keyword matching coefficients; The analysis conditions include a first analysis condition and a second analysis condition. The first analysis condition is that the proportion of risk images is greater than or equal to the preset proportion of risk images or the mean of the sensitive feature distribution is greater than or equal to the preset mean of the sensitive feature distribution. The second analysis condition is that the proportion of risk images is less than the preset proportion of risk images and the mean of the sensitive feature distribution is less than the preset mean of the sensitive feature distribution. The desensitization processing unit responds to the first analysis condition and determines that the desensitization method is to select a reference number of sensitive features in each feature combination for fuzzy processing; The desensitization processing unit responds to the second analysis condition and determines that the desensitization method is to perform fuzzy processing on all sensitive features; The desensitization method of the desensitization processing unit response is to select a reference number of sensitive features in each feature combination for fuzzy processing, wherein, When performing a combined analysis on a single sensitive feature in a single evidence image, the sensitive feature is recorded as a target sensitive feature, and other sensitive features in the evidence image other than the target sensitive feature are recorded as reference sensitive features. The set of reference sensitive features and the target sensitive features whose feature correlation with the target sensitive feature is greater than the preset feature correlation is recorded as a feature combination; Continue to perform combined analysis on sensitive features that are not recorded in the feature combination until all sensitive features are recorded in the feature combination. Randomly select a reference number of sensitive features in a single feature combination for fuzzy processing. The reference number is positively correlated with the number of sensitive features corresponding to the single feature combination. The sensitive features include sensitive areas and sensitive keywords.

2. The analysis and evaluation system for password security according to claim 1, characterized in that: The keyword matching unit is responsive to the evidence image type to determine a keyword matching method for the evidence image; The keyword matching unit responds that the evidence image type is a text evidence image, and determines the keyword matching method as determining the matching keyword based on the number of times the combination is used; The keyword matching unit responds that the evidence image type is a scene evidence image, and determines the keyword matching method as determining the matching keyword based on the feature matching degree; The evidence image types include text evidence images and scene evidence images.

3. The analysis and evaluation system for password security according to claim 2, characterized in that: The keyword matching unit determines matching keywords according to the number of times the combination is used, including: For a text evidence image, the text evidence image is recorded as a target text evidence image, and other text evidence images other than the target text evidence image are recorded as reference text evidence images, and each similar image corresponding to the target text evidence image is detected, and matching analysis is performed on each similar image; When performing matching analysis on a single similar image, the similar image is recorded as the reference similar image, the same keywords in the target text evidence image and the reference similar image are recorded as reference same keywords, the other keywords in the target text evidence image other than the reference keywords are recorded as keywords to be analyzed, and the keywords in the uploaded data that are used in combination with the single keyword to be analyzed more than the preset number of times are recorded as collocation keywords; Each keyword in the target text evidence image and the collocation keyword corresponding to each keyword to be analyzed are recorded as the matching keyword of the target text evidence image; The similar image is a reference text evidence image in which the number of identical keywords with the target text evidence image is greater than a preset number of identical keywords and the distribution similarity of identical keywords is greater than a preset distribution similarity of identical keywords.

4. The analysis and evaluation system for password security according to claim 3, characterized in that: The image selection unit responds to the set conditions to determine the image selection method of the secret review text; The setting condition of the image selection unit response is that the proportion of the influencing keywords is greater than or equal to the preset proportion of the influencing keywords or the influencing factor is greater than or equal to the preset influencing factor, and the image selection method is to select the evidence image according to the characteristic keywords; The setting condition of the image selection unit response is that the proportion of the influencing keywords is less than the preset proportion of the influencing keywords and the influencing factor is less than the preset influencing factor, and the image selection method is to select the evidence image according to the keyword matching degree; The characteristic keywords are keywords in the influence keyword group whose sub-influence factors are greater than the preset sub-influence factors.

5. The analysis and evaluation system for password security according to claim 4, characterized in that: The image selection unit determines the sub-influence factor according to the disorder coefficient, wherein, If the disorder coefficient is greater than or equal to the preset disorder coefficient, the sub-influence factor is determined according to the influence mean; If the disorder coefficient is less than the preset disorder coefficient, the sub-influence factor is determined according to the paragraph length and the number of associations.

6. The analysis and evaluation system for password security according to claim 4, characterized in that: The risk determination unit responds to the evidence image type to determine the image risk value determination method; The risk determination unit corresponds to the scene evidence image and determines the image risk value according to the area and number of sensitive areas; The risk determination unit corresponds to the text evidence image and determines the image risk value based on the number of sensitive keywords.

7. The analysis and evaluation system for password security according to claim 1, characterized in that: The desensitization processing unit determines the feature correlation degree according to the sensitive features; If the sensitive feature is a sensitive keyword, the feature relevance is determined based on the collocation coefficient and the interval reference value; If the sensitive feature is a sensitive area, the feature association degree is determined based on the device association degree.

8. The analysis and evaluation system for password security according to claim 7, characterized in that: The auxiliary selection unit determines a supplementary selection method for the evidence image according to the image correlation; If the image correlation is greater than or equal to the preset image correlation, the supplementary selection method is to select supplementary images in descending order of feature correlation coefficients; If the image relevance is less than the preset image relevance, the supplementary selection method is to select supplementary images in descending order of keyword matching coefficients.

Images