Sensitive information authorization display system based on face recognition

Abstract

The present invention discloses a sensitive information authorization display system based on face recognition, relates to the field of face recognition technology, and is used to solve the problem that traditional face recognition systems usually rely on fixed models and static image processing strategies, and cannot effectively cope with changes in different environmental conditions. Traditional systems often ignore the impact of network status on face recognition systems, especially when network bandwidth is limited, image transmission and processing may be restricted, affecting the recognition result. By recording historical data of face recognition images at each moment, the system can perform comparative analysis based on time series data to find out whether there are significant changes in the image. The system can dynamically adjust the processing method according to different situations, so that the face recognition system can maintain high stability and accuracy in different environments and network conditions, thereby ensuring the long-term stable operation of the system.

Description

Technical Field

[0001] The present invention relates to the field of face recognition technology, and more specifically, to a sensitive information authorization display system based on face recognition. Background Art

[0002] The information authorization display system based on face recognition combines multiple technologies such as face recognition technology, information authorization management, information display technology, security protection measures and deep learning. The advancement of these technologies has promoted the improvement of intelligent security, convenience and user experience, and has broad application prospects in security, finance, medical and other fields. With the continuous maturity of technology and the deepening of application, the information authorization display system based on face recognition will play a more important role in the future intelligent society.

[0003] The prior art has the following deficiencies:

[0004] Traditional face recognition systems usually rely on fixed models and static image processing strategies, which cannot effectively cope with changes in different environmental conditions. Traditional systems often ignore the impact of network status on face recognition systems, especially when network bandwidth is limited, image transmission and processing may be restricted, affecting recognition results. By monitoring the network status of the device in real time in the background monitoring module and selecting appropriate processing methods according to network conditions (such as using simpler grayscale processing when the network is poor), the system can maintain good recognition performance even when the network is poor. Summary of the invention

[0005] In order to overcome the above-mentioned defects of the prior art, an embodiment of the present invention provides a sensitive information authorization display system based on face recognition to solve the problems raised in the above-mentioned background technology.

[0006] To achieve the above object, the present invention provides the following technical solutions:

[0007] The sensitive information authorization display system based on face recognition includes: a data acquisition module, an abnormality judgment module and a background monitoring module, and the signal connection between the modules;

[0008] Data collection module: records the historical data of face recognition images at each moment and obtains the current weather data;

[0009] Abnormal judgment module: calculates the image matching degree based on the historical data of face recognition images and weather data at each moment to determine whether the current image is abnormal;

[0010] Background monitoring module: obtains the complexity of the background in the image and the network status of the current face recognition device;

[0011] Differentiation region module: Based on the complexity of the background and the current network status, grayscale processing or super-pixel processing is used to distinguish the face from the external area in the image.

[0012] In a preferred embodiment, the data acquisition module includes the following contents:

[0013] The system uses facial recognition equipment to regularly collect images of the current facial recognition. Each image is accompanied by a timestamp, which records the specific time when the image was collected and stores historical data. The system obtains current real-time weather data through the API.

[0014] In a preferred embodiment, the abnormality judgment module includes the following contents:

[0015] Use the mean square error to judge the image and the face recognition image at the same time in the historical data, compare the similarity between the two images, and design a weight coefficient based on weather data to obtain the matching degree and design a threshold. If the threshold is exceeded, the background is abnormal.

[0016] The image is grayed out, divided into N equal pixel points, and the gray value of each pixel is recorded. The mean square error (MSE) is calculated between the current image and the image at the same time in the historical data. The mean square error (MSE) is calculated as follows: Where: I1(i) is the grayscale value of the historical data of the face recognition image at the same time at the pixel position i, I2(i) is the grayscale value of the current image at the pixel position i, and N is the total number of pixels in the image;

[0017] The MSE value is normalized and expressed as Where: MSE max is the maximum value of MSE, MSE mim is the minimum value of MSE.

[0018] To quantify the current real-time weather data obtained, modeling needs to be based on the following two factors: First, the change of weather types needs to be considered, and an influencing factor needs to be defined for different weather types. The magnitude of weather changes can be quantified through meteorological data;

[0019] Weather Impact is defined as the degree of weather change, and the formula is: Weather Impact = Weather Change Index × Weather Sensitivity Factor, where: Weather Change Index is calculated based on the change in temperature, humidity, or other meteorological parameters, and Weather Sensitivity Factor is a coefficient adjusted according to the characteristics of the background image;

[0020] Background Difference Score=α·MSE normalized +β·Weather Impact, where: Background Difference Score: measures the background difference of the image, MSE n0rmalized : is the normalized MSE value, Weather Impact: indicates the impact of weather changes on the image background, α and β are weight coefficients;

[0021] Finally, the background difference score Background Difference Score is compared with the set threshold to determine whether a background abnormality has occurred. If the Background Difference Score exceeds a certain threshold, it means that an abnormality has occurred in the background.

[0022] In a preferred embodiment, the background monitoring module includes the following contents:

[0023] Analyze the background complexity of the image through the color histogram: convert the image to RGB color space, calculate the histogram of each color channel separately, and use the RGB channels to calculate and normalize the entropy value of the histogram. The calculation formula is as follows: Where P(i) is the probability of pixel value i in each color channel, and H is the entropy value;

[0024] According to the size of the entropy value, a threshold can be set to judge the background complexity of the image. A suitable threshold T can be obtained through experiments or data analysis. If the entropy value H of the image exceeds this threshold, the background of the image can be considered complex; if the entropy value is lower than the threshold, the background can be considered simple.

[0025] Quantify the network status of the device and set thresholds for evaluation. Evaluate the key indicators of bandwidth, latency, packet loss rate and jitter, calculate according to different threshold intervals and weights, and finally get a comprehensive score. The formula is: Network status score = bandwidth score × 30% + latency score × 30% + packet loss rate score × 20% + jitter score × 20%;

[0026] According to the size of the network status score, a threshold can be set to judge the quality of the network. A suitable threshold can be obtained through experiments or data analysis. If the network status score exceeds this threshold, the network status quality is considered to be high; if the entropy value is lower than the threshold, the network status quality is considered to be low.

[0027] In a preferred embodiment, the differentiation region module includes the following:

[0028] After receiving the background complex entropy value of the image and the current network state, the differentiation region module defines the background complex entropy value of the image and the current network state as input variables and divides them into different fuzzy sets respectively;

[0029] The processing model is defined as the output variable, which is divided into fuzzy sets;

[0030] Fuzzy rules are formulated to describe the background complex entropy value of the image and the impact of the current network state definition on the processing model.

[0031] The technical effects and advantages of the sensitive information authorization display system based on face recognition of the present invention are as follows:

[0032] By recording the historical data of face recognition images at each moment, the system can perform comparative analysis based on time series data. The current image can be compared with the historical data to find out whether there are significant changes in the image. This comparison can help the system identify anomalies in the image, such as light changes, occlusion, or face recognition failure. Weather data as supplementary information: Weather conditions (such as light, rain, snow, haze, etc.) can affect image quality or face recognition accuracy. By combining real-time weather data, the system can adjust the judgment criteria more reasonably. For example, bad weather may cause blurred or low-quality images. At this time, a loose matching standard can be set to avoid misjudgment as anomalies. The system can dynamically adjust the processing method according to different situations, so that the face recognition system can maintain high stability and accuracy in different environments and network conditions. Whether it is in the case of changes in lighting, changes in background complexity, or fluctuations in network status, the system can respond flexibly to ensure the long-term stable operation of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] Figure 1 This is a schematic diagram of the structure of the sensitive information authorization display system based on face recognition of the present invention. DETAILED DESCRIPTION

[0034] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0035] The present invention is collected.

[0036] Example 1

[0037] The invention discloses a sensitive information authorization display system based on face recognition, comprising: a data acquisition module, an abnormality judgment module, a background monitoring module and a differentiation area module, and the modules are signal-connected.

[0038] Data collection module: records the historical data of face recognition images at each moment and obtains the current weather data;

[0039] Abnormal judgment module: calculates the image matching degree based on the historical data of face recognition images and weather data at each moment to determine whether the current image is abnormal;

[0040] Background monitoring module: obtains the complexity of the background in the image and the network status of the current face recognition device;

[0041] Differentiation region module: Based on the complexity of the background and the current network status, grayscale processing or super-pixel processing is used to distinguish the face from the external area in the image.

[0042] The functions of each module are as follows:

[0043] In the data collection module, the goal of this stage is to record the historical data of face recognition images at each moment and obtain the current weather data. The specific contents include:

[0044] The system uses the face recognition device to collect images of the current face recognition at a regular interval (for example, every second or every minute). Each image should be accompanied by a timestamp to record the specific time when the image was collected to facilitate subsequent historical data comparison and storage.

[0045] Get the current real-time weather data through the API and the impact of lighting conditions on image changes: sunny and cloudy weather will result in different light intensities, affecting the brightness and shadows of the image.

[0046] In the abnormality judgment module, the image matching degree is calculated based on the historical data of face recognition images and weather data at each moment to determine whether the current image is abnormal. The specific contents include:

[0047] The mean square error is used to judge the image and the face recognition image at the same time in the historical data, and the similarity between the two images is compared. The weight coefficient is designed in combination with the weather data to obtain the matching degree and a threshold is designed. If the threshold is exceeded, the background is abnormal.

[0048] It should be noted that the mean square error is a common standard for measuring the error of prediction models in statistics and machine learning. It is an important indicator for evaluating the accuracy of regression models. It represents the average of the squares of the difference between the model prediction value and the actual value. The smaller the MSE, the smaller the difference in grayscale values ​​between the current image and the historical image, and the higher the similarity between the images. Ideally, when two images are exactly the same, the MSE is 0. The larger the MSE, the more significant the difference in grayscale values ​​between the two images, that is, the greater the difference between the two images.

[0049] The image is grayed out, divided into N equal pixel points, and the gray value of each pixel is recorded. The mean square error (MSE) is calculated between the current image and the image at the same time in the historical data. The mean square error (MSE) is calculated as follows: Where: I1(i) is the grayscale value of the historical data of the face recognition image at the same time at the pixel position i, I2(i) is the grayscale value of the current image at the pixel position i, and N is the total number of pixels in the image.

[0050] For example: suppose there are two 3x3 grayscale images (9 pixels), where:

[0051] The gray value I2(i) of the current image is:

[0052] The gray value I1(i) of the historical image is:

[0053]

[0054] To facilitate subsequent processing, the MSE value is normalized, and the normalized MSEZ value will be between 0 and 1. Where: MSE max is the maximum value of MSE, MSE min is the minimum value.

[0055] The current real-time weather data acquired is quantified to facilitate subsequent weighted processing. Modeling needs to be based on the following factors: First, the weather type changes need to be considered: from sunny to cloudy, rainy or snowy, the background may change significantly. An impact factor can be defined for different weather types, such as sunny = 0, cloudy = 0.2, rainy = 0.5, snowy = 0.7. Weather change range: It can be quantified through meteorological data, such as changes in temperature, humidity, wind speed, etc. If some weather changes are more drastic (such as a sudden change from sunny to heavy rain), its impact on the background will be greater.

[0056] Weather Impact is defined as the degree of weather change, and the formula is: Weather Impact = Weather Change Index × Weather Sensitivity Factor, where: Weather Change Index is calculated based on the change in temperature, humidity, or other meteorological parameters, and Weather Sensitivity Factor is a coefficient adjusted according to the characteristics of the background image.

[0057] Background Difference Score=α·MSE normalized+β·Weather Impact, where: Background Difference Score: This score is used to measure the background difference of the image, MSE normalized is the normalized MSE value. Weather Impact: It indicates the degree of influence of weather changes on the image background. For example, a change from sunny to rainy will cause significant changes in background illumination, reflection, color, etc., thus affecting the recognizability of the background. Weather impact can be quantified by the change in weather type or the magnitude of weather change. α and β are weight coefficients that determine MSE. mormalized The relative importance of the image similarity and weather impact in the final score can be controlled by adjusting these two coefficients to control the impact of image similarity and weather factors on background differences.

[0058] Finally, the background difference score is compared with the set threshold to determine whether a background anomaly has occurred. The threshold can be set based on the background difference score of historical data or automatically optimized through a machine learning algorithm. If the Background Difference Score exceeds a certain threshold, it indicates that an abnormality may have occurred in the background, such as abrupt changes in objects or weather in the background, or decreased image quality.

[0059] For example:

[0060] Suppose there is a scene that transitions from sunny to heavy rain, the MSE of the image normalized The value is 0.85, indicating that the similarity with the reference image is low. At the same time, the weather change leads to a significant change in the scene background, and the Weather Impact calculation value is 0.6 (derived from the combination of weather type changes and factors such as temperature and humidity). Assuming that α = 0.7 and β = 0.3 are set, the background difference score is:

[0061] Background Difference Score=α·MSE normalized +β·Weather Impact=0.7×0.85+0.3×0.6=0.595+0.18=0.775

[0062] If the threshold is set to 0.8, the background difference score is less than the threshold, indicating that there is no obvious abnormality in the background; if the threshold is set to 0.7, it means that there is an obvious abnormality in the background.

[0063] In the background monitoring module, the complexity of the background in the image and the network status of the current face recognition device are obtained, including:

[0064] The background complexity of the image is analyzed by the color histogram. The color histogram is a graphical representation of the frequency distribution of each color in the image. The color distribution and complexity of the image can be determined. First, the color channel of the image is extracted: the image is converted to the RGB color space, and the histogram of each color channel (red R, green G, blue B) is calculated separately, showing the distribution of pixel values ​​of each color in the image. The RGB channels are used for calculation separately. In order to make the histograms of different images comparable, the histograms are usually normalized. The normalized histogram can make the frequency value of each color channel between 0 and 1, avoiding the frequency difference caused by different image sizes, and calculating the entropy value of the histogram. The entropy value is a measure of the amount of information, which can reflect the distribution complexity of the histogram. The higher the entropy value, the more uniform the color distribution and the more complex the background; the lower the entropy value, the uneven color distribution and the simpler the background. The entropy value is calculated as follows: Among them, P(i) is the probability that the pixel value in each color channel is i, and H is the entropy value.

[0065] According to the size of the entropy value, a threshold can be set to judge the background complexity of the image. A suitable threshold T can be obtained through experiments or data analysis. If the entropy value H of the image exceeds this threshold, then the background of the image can be considered to be relatively complex; if the entropy value is lower than the threshold, then the background can be considered to be relatively simple.

[0066] Quantifying the network status of devices and setting thresholds for evaluation can be achieved through the following key indicators. Common network status indicators include bandwidth, latency, packet loss rate, and jitter. Through real-time monitoring and analysis of these indicators, the quality of the network can be judged.

[0067] Bandwidth: refers to the maximum transmission speed of a network connection, usually expressed in Mbps (megabits per second) or Gbps (gigabits per second). When the bandwidth is higher, the device can transfer more data and the network status is generally better.

[0068] Latency: refers to the round-trip time for data to travel from the source device to the destination device, usually expressed in milliseconds (ms). High latency can lead to a poor network experience, especially in real-time communications or online games.

[0069] Packet loss rate: Indicates the proportion of data packets lost during data transmission. A high packet loss rate usually indicates an unstable network, which may be caused by network congestion, equipment failure, etc.

[0070] Jitter: refers to the fluctuation or change of network delay. In real-time communications (such as video conferencing, VoIP), excessive jitter will affect the audio and video quality.

[0071] Network throughput: refers to the amount of data actually transmitted per unit time, usually measured in Mbps. Throughput is affected by bandwidth, latency, and packet loss, and is an important indicator for measuring network performance.

[0072] The calculation is based on different threshold intervals and weights, and finally a comprehensive score is obtained. The formula is: Network status score = bandwidth score × 30% + delay score × 30% + packet loss rate score × 20% + jitter score × 20%

[0073] According to the size of the network status score, a threshold can be set to judge the quality of the network. A suitable threshold can be obtained through experiments or data analysis. If the network status score exceeds this threshold, then the network status quality can be considered high; if the entropy value is lower than the threshold, then the network status quality can be considered low.

[0074] In the differentiation region module, grayscale processing or superpixel processing is used to distinguish the face and external areas in the image according to the background complexity entropy value and the current network status. The specific contents include:

[0075] After receiving the background complex entropy value of the image and the current network state, the differentiation region module defines the background complex entropy value of the image and the current network state as input variables and divides them into different fuzzy sets.

[0076] For example, "Low", "Medium", and "High" represent the background complex entropy value, and "Low", "Medium", and "High" represent the current network status.

[0077] Define the processing model as the output variable and partition it into fuzzy sets, for example, "Grayscale", "Superpixel" for the processing model.

[0078] Formulate a set of fuzzy rules to describe the impact of different input variables on output variables. The definition of rules can be based on professional knowledge or obtained through data analysis and experiments. For example:

[0079] The background complex entropy value is marked as H, the current network state is marked as G, and the processing model is marked as Model.

[0080] Rule 1:IF(H is Low)AND(G is Low)THEN(Model is Grayscale)

[0081] Rule 2:IF(H is High)AND(G is High)THEN(Model is Superpixel) ...

[0083] Perform fuzzy reasoning based on fuzzy rules and determine the solution for processing the model.

[0084] It should be noted that the division of fuzzy sets can be adjusted according to actual conditions. For example, although this embodiment takes three fuzzy sets as examples, the background complex entropy value, the current network status and the processing model can actually be divided into more than three sets to facilitate better precise adjustment according to different temperatures.

[0085] The above formulas are all dimensionless and numerical calculations. The formula is a formula for the most recent real situation obtained by collecting a large amount of data and performing software simulation. The preset parameters in the formula are set by technicians in this field according to actual conditions.

[0086] The above embodiments may be implemented in whole or in part by software, hardware, firmware or any other combination. When implemented by software, the above embodiments may be implemented in whole or in part in the form of a computer program product.

[0087] Those of ordinary skill in the art will appreciate that the modules and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application of the technical solution and the invention constraints. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0088] In addition, each functional module in each embodiment of the present application may be integrated into one processing module, or each module may exist physically separately, or two or more modules may be integrated into one module.

[0089] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

[0090] Finally: The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. Sensitive information authorization display system based on face recognition, characterized in that: include: Data acquisition module, abnormality judgment module and background monitoring module, and signal connection between each module; Data collection module: records the historical data of face recognition images at each moment and obtains the current weather data; Abnormal judgment module: calculates the image matching degree based on the historical data of face recognition images and weather data at each moment to determine whether the current image is abnormal; Background monitoring module: obtains the complexity of the background in the image and the network status of the current face recognition device; Differentiation region module: Based on the complexity of the background and the current network status, grayscale processing or super-pixel processing is used to distinguish the face from the external area in the image.

2. The sensitive information authorization display system based on face recognition according to claim 1 is characterized in that: Record the historical data of face recognition images at each moment and obtain the current weather data: The system uses facial recognition equipment to regularly collect images of the current facial recognition. Each image is accompanied by a timestamp, which records the specific time when the image was collected and stores historical data. The system obtains current real-time weather data through the API.

3. The sensitive information authorization display system based on face recognition according to claim 2 is characterized in that: The image matching degree is calculated based on the historical data of face recognition images and weather data at each moment to determine whether the current image is abnormal: Use the mean square error to judge the image and the face recognition image at the same time in the historical data, compare the similarity between the two images, and design a weight coefficient based on weather data to obtain the matching degree and design a threshold. If the threshold is exceeded, the background is abnormal. The image is grayed out, divided into N equal pixel points, and the gray value of each pixel is recorded. The mean square error (MSE) is calculated between the current image and the image at the same time in the historical data. The mean square error (MSE) is calculated as follows: Where: I1(i) is the grayscale value of the historical data of the face recognition image at the same time at the pixel position i, I2(i) is the grayscale value of the current image at the pixel position i, and N is the total number of pixels in the image; The MSE value is normalized and expressed as Where: MSE max is the maximum value of MSE, MSE min is the minimum value of MSE; To quantify the current real-time weather data obtained, modeling needs to be based on the following two factors: First, the change of weather types needs to be considered, and an influencing factor is defined for different weather types. The magnitude of weather changes is quantified through meteorological data; Weather Impact is defined as the degree of weather change, and the formula is: Weather Impact = Weather Change Index × Weather Sensitivity Factor, where: Weather Change Index is calculated based on the magnitude of temperature and humidity changes, and Weather Sensitivity Factor is a coefficient adjusted according to the characteristics of the background image; Background Difference Score= α MSE normalized +β·Weather Impact , where: Background Difference Score: measures the degree of background difference of the image, MSE normalized : is the normalized MSE value, Weather Impact: indicates the degree of influence of weather changes on the image background, α and β are weight coefficients; Finally, the background difference score is compared with the set threshold to determine whether a background abnormality occurs.

4. The sensitive information authorization display system based on face recognition according to claim 3 is characterized in that: Obtain the complexity of the background in the image and the network status of the current face recognition device; Analyze the background complexity of the image through the color histogram: convert the image to RGB color space, calculate the histogram of each color channel separately, and use the RGB channels to calculate and normalize the entropy value of the histogram. The calculation formula is as follows: Where P(i) is the probability of pixel value i in each color channel, and H is the entropy value; According to the size of the entropy value, a threshold can be set to judge the background complexity of the image. A suitable threshold T can be obtained through experiments or data analysis. If the entropy value H of the image exceeds this threshold, the background of the image can be considered complex; if the entropy value is lower than the threshold, the background can be considered simple. Quantify the network status of the device and set thresholds for evaluation. Evaluate the key indicators of bandwidth, latency, packet loss rate and jitter, calculate according to different threshold intervals and weights, and finally get a comprehensive score. The formula is: Network status score = bandwidth score × 30% + latency score × 30% + packet loss rate score × 20% + jitter score × 20%; According to the size of the network status score, a threshold can be set to judge the quality of the network. A suitable threshold can be obtained through experiments or data analysis. If the network status score exceeds this threshold, the network status quality is considered to be high; if the entropy value is lower than the threshold, the network status quality is considered to be low.

5. The sensitive information authorization display system based on face recognition according to claim 4 is characterized in that: Depending on the complexity of the background and the current network status, choose to use grayscale processing or superpixel processing to distinguish the face from the external area in the image: After receiving the background complex entropy value of the image and the current network state, the differentiation region module defines the background complex entropy value of the image and the current network state as input variables and divides them into different fuzzy sets respectively; The processing model is defined as the output variable, which is divided into fuzzy sets; Fuzzy rules are formulated to describe the background complex entropy value of the image and the impact of the current network state definition on the processing model.

Images