Video monitoring system for medical health big data privacy management

Abstract

The invention discloses a video monitoring system for medical health big data privacy management, and relates to the technical field of video monitoring, the system comprises the following components: an image acquisition module, an environment and privacy management module, a data monitoring module, a face recognition module and a privacy display control module; through combination of a high-resolution and low-illumination camera of the image acquisition module and a scene analysis technology, automatic adaptation to different medical environments and illumination conditions is realized, and meanwhile, the environment and privacy management module automatically sets and adjusts monitoring strategies and camera parameters according to privacy levels of different areas, so that the safety of the environment and privacy management module is improved. The definition and availability of a monitoring picture are ensured, the labor cost is effectively reduced, in addition, the system is excellent in the aspect of protecting the privacy of the patient, acquisition of the privacy of the patient by unauthorized personnel is strictly limited through the technical means of reducing the monitoring sensitivity of a ward and other privacy areas, fuzzy processing and the like, and the safety of the patient is improved. And the privacy protection level of the medical place is greatly improved.

Description

Technical Field

[0001] The present invention relates to the field of video surveillance technology, and in particular to a video surveillance system for privacy management of medical health big data. Background Art

[0002] In the medical environment, video surveillance systems play a vital role. They not only help to ensure the safety of medical facilities, but also help to manage medical affairs and to a certain extent assist medical services. With the continuous advancement of medical technology and the rapid growth of medical data, the demand for video surveillance systems is also increasing. Especially in the privacy management of medical and health big data, video surveillance systems need to be able to provide more efficient and intelligent monitoring solutions.

[0003] However, traditional video surveillance systems in medical facilities have obvious shortcomings. On the one hand, these systems are often difficult to automatically adapt to different medical environments and lighting conditions, and require frequent manual adjustment of camera parameters to obtain clear images, which not only increases labor costs, but may also result in poor image quality due to failure to adjust in time, affecting the monitoring effect. On the other hand, in terms of privacy protection, traditional systems find it difficult to automatically and flexibly adjust the monitoring range and accuracy according to the privacy sensitivity of different areas. For example, in ward areas, excessive monitoring may infringe on patient privacy, while in key areas such as data centers, insufficient monitoring may not be able to effectively protect the security of important data and facilities. In addition, traditional systems also have shortcomings in identity recognition accuracy and real-time performance, making it difficult to effectively prevent unauthorized personnel from accessing medical and health big data. At the same time, in the processing of privacy areas involving monitoring images, traditional systems lack flexibility and intelligence, and cannot perform targeted display control according to different personnel roles and permissions.

[0004] In view of the above problems, it is necessary to optimize the existing video surveillance system for privacy management of medical and health big data. By introducing technical means such as image acquisition, environment and privacy management, data monitoring, facial recognition and privacy display control, it realizes automatic adaptation to different medical area environments, adjustment of privacy protection strategies and real-time monitoring of facial recognition. Therefore, it is of great significance to develop a video surveillance system for privacy management of medical and health big data that can comprehensively realize the above characteristics. Summary of the invention

[0005] The purpose of the present invention is to make up for the shortcomings of the prior art and provide a video monitoring system for privacy management of medical and health big data. It can automatically adapt to different medical environments and lighting conditions through the high-resolution, low-light camera in the image acquisition module, combined with the scene analysis capability. At the same time, the environment and privacy management module automatically adjusts the camera parameters according to the privacy level and monitoring strategy to ensure the acquisition of clear and usable video images. The data monitoring module encrypts the video image data and displays the processed images in real time to facilitate viewing and supervision by monitoring personnel. The facial recognition module adopts a facial recognition algorithm to realize dynamic facial recognition of people entering the medical data management area to improve security. The privacy display control module flexibly controls the display degree of the privacy area according to the role and authority of the personnel to effectively protect the privacy of patients.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions: a video monitoring system for privacy management of medical health big data, the system comprising the following components: an image acquisition module, an environment and privacy management module, a data monitoring module, a facial recognition module and a privacy display control module; The image acquisition module is composed of multiple cameras distributed in various areas of the medical site. The cameras have high resolution and low illumination characteristics, and have scene analysis capabilities. They can automatically analyze the medical environment and lighting conditions in real time, and transmit the analysis results to subsequent modules, providing the original video image data source for the entire monitoring system; The environment and privacy management module formulates monitoring strategies for each area based on the privacy level information of different areas. For areas with high privacy requirements, the monitoring scope is limited to the ward entrance and public activity area to avoid excessive monitoring of the patient's private rest area in the ward. At the same time, the resolution of image acquisition is reduced so that the collected images meet the security monitoring requirements while not causing the patient's privacy details to be exposed. For key areas, considering their impact on medical data security and system operation stability, the area is monitored in all directions and the resolution of image acquisition is improved. At the same time, the intelligent recognition function is turned on to provide real-time warnings for abnormal personnel and behaviors through image recognition technology. Once unauthorized entry or abnormal behavior is found, an alarm is issued in time and corresponding measures are taken. For monitoring strategies, the camera parameters are finely adjusted according to the pre-constructed lighting conditions and parameter mapping relationship table. The adjusted camera parameters cover multiple aspects, including but not limited to focal length, aperture, sensitivity and white balance. The data monitoring module is responsible for receiving and processing video image data after environmental and privacy management, using communication protocols and data verification mechanisms to ensure stable and complete data transmission, and converting data into a format suitable for medical and health big data analysis through algorithms. At the same time, encryption algorithms are used to encrypt the parts involving patient privacy, and the processed video images are displayed in real time through the multi-screen split display function, and specific screens can be enlarged, reduced, and paused according to the operating instructions of the monitoring personnel; The facial recognition module is set at the entrance of the medical data management area and key locations around it. Before the system is activated, facial feature data of legal management personnel are collected and stored as reference data. When a person is detected entering the area, a high-precision image acquisition device is used to collect their facial images in real time. The real-time image is compared with the reference data through a facial recognition algorithm, and the factors of various facial feature points are comprehensively considered. If the comparison is successful, it is determined that the person is authorized to enter. If it fails, it is determined that the person is unauthorized and an alarm is issued while recording relevant information; The privacy display control module uses image recognition technology to identify the content involving privacy areas in the monitoring screen, and uses image processing algorithms to dynamically block it, making part of the content invisible. At the same time, based on the pre-stored permission information corresponding to different personnel roles, when the monitoring personnel views the monitoring screen, they determine whether to display and the degree of display according to their role permissions, thereby achieving differentiated privacy management.

[0007] Furthermore, the image acquisition module automatically performs real-time analysis of the medical environment and lighting conditions. For the medical environment analysis, the formula Determine whether the area is a specific medical area, where: Represents the area matching score, which is used to measure the matching degree between the current camera area and various predefined medical areas. is the number of features used for region recognition, It is The weight coefficient of regional features, It is The matching degree of regional features is calculated using the formula Analyze the personnel turnover, including: It represents the personnel flow index, which is used to reflect the density and changes of personnel flow in the current region. is the number of human targets tracked within a certain time interval, It is The frequency with which a person target appears in a specific time period. is the area covered by the current camera monitoring range. is the coefficient of change of personnel turnover, is the variance of personnel distribution. For lighting conditions, the formula is used Analyze the light intensity, where Indicates the average light intensity currently perceived by the camera. is the number of image pixels selected to measure light intensity, It is The light intensity value of the selected pixel point is calculated using the formula Analyze the light color temperature, where Indicates the color temperature of the current light. is the number of image pixels selected for analyzing the light color temperature, It is The color component value of the selected pixel in the RGB color space, It is the color temperature conversion coefficient corresponding to the RGB color component value.

[0008] Furthermore, the environment and privacy management module formulates a monitoring strategy for each area based on the privacy level information of different areas, and its algorithm formula is: ,in, Indicates the index value of the monitoring strategy finally selected, corresponding to a specific strategy in the set of pre-stored different monitoring strategies. is the privacy level parameter, is the data security weight coefficient, is the privacy protection weight coefficient.

[0009] Furthermore, the environment and privacy management module fine-tunes the camera parameters for the monitoring strategy according to the pre-built lighting condition and parameter mapping relationship table, and the calculation formula is: ,in, is the adjusted camera parameter value, including focal length, aperture, sensitivity and white balance, It is the basic camera parameter value based on the current regional monitoring strategy. is the monitoring strategy influence coefficient, is the influence coefficient of lighting conditions, is the monitoring strategy index value obtained by the monitoring strategy selection formula, is the average value of all monitoring strategy index values, and They are the reference light intensity value and the current light intensity value respectively.

[0010] Furthermore, the data monitoring module converts the data into a format suitable for medical and health big data analysis through an algorithm, and the algorithm formula is: ,in, Indicates the converted video image data format. It is the original video image data format, that is, the unconverted data format received from the environment and privacy management module. It is a specific data specification required for medical and health big data analysis. is the corresponding data specification in the original data format, is the data format adjustment factor, is a key data indicator in the original data format, It is the corresponding key data indicator required for medical and health big data analysis.

[0011] Furthermore, the data monitoring module uses an encryption algorithm to encrypt the part involving patient privacy, and the algorithm formula of the encryption algorithm is: ,in, Indicates the encrypted ciphertext, that is, the data after encryption. Indicates the key For plain text To perform encryption operations, Is the plaintext data that needs to be encrypted. is the symmetric encryption key, is the key obfuscation coefficient, Is the key The hash value of .

[0012] Furthermore, the facial recognition module uses a high-precision image acquisition device to collect facial images in real time, and compares the real-time image with the reference data through a facial recognition algorithm, the formula of which is: ,in, Represents the facial matching score, which is an indicator that comprehensively evaluates the matching degree between the real-time acquired facial image and the pre-stored benchmark data. Less than threshold , then it is judged as a comparison failure, that is, it is considered to be an unauthorized person. If Greater than or equal to , then the comparison is judged to be successful, that is, it is considered to be an authorized person, It is the first facial image collected in real time. The eigenvalues ​​of feature points, It is the first one in the pre-stored benchmark data. The eigenvalues ​​of feature points, It is The weight coefficient of feature points, is the total number of selected facial feature points, is the action influence coefficient, It is The variation range of a feature point under the natural movement of the person.

[0013] Furthermore, the privacy display control module uses image recognition technology to identify the content involving the privacy area in the monitoring screen. Specifically, through the formula Accurately identify the privacy area in the monitoring screen, including: It represents the privacy area identification index, which is an indicator for comprehensively evaluating whether a certain area in the monitoring image is a privacy area. Above the set threshold , then the area is determined to be a privacy area. If Below threshold , then the area is determined not to be a privacy area. It is The weight coefficients of the image content elements, It is the first The existence probability of the image content elements, is the total number of image content elements selected, is the area influence coefficient, It is The area ratio of the area where the image content element is located is the ratio of the area where the image content element is located to the area of ​​the entire monitoring screen.

[0014] Furthermore, the privacy display control module determines whether to display and the degree of display according to the pre-stored authority information corresponding to different personnel roles when the monitoring personnel view the monitoring screen according to their role authority. The judgment formula for the degree of display is: ,in, Indicates the final display degree of the privacy area. It is the personnel role permission parameter. is the initial occlusion degree of the privacy area, is the authority adjustment coefficient, It is the role code of the person currently viewing the surveillance screen. The code can be used to quickly identify the person's role type. is the default role encoding.

[0015] Compared with the prior art, this video surveillance system for medical health big data privacy management has the following beneficial effects: 1. The present invention realizes automatic adaptation to different medical environments and lighting conditions through the combination of high-resolution, low-light cameras and scene analysis technology of the image acquisition module. At the same time, the environment and privacy management module automatically sets and adjusts the monitoring strategy and camera parameters according to the privacy level of different areas, ensuring the clarity and availability of the monitoring images and effectively reducing labor costs. In addition, the system performs well in protecting patient privacy. By reducing the monitoring sensitivity of private areas such as wards and using technical means such as fuzzy processing, it strictly limits the access of unauthorized personnel to patient privacy, greatly improving the level of privacy protection in medical facilities.

[0016] 2. The present invention realizes dynamic facial recognition of people entering the medical data management area through a facial recognition module, improves the accuracy and practicality of recognition, and effectively prevents unauthorized entry, thereby ensuring that only legitimate managers can access medical and health big data. In addition, the privacy display control module flexibly controls the display degree of the privacy area according to the role authority of the monitoring personnel, which not only protects the patient's privacy, but also meets the needs of different personnel to view the monitoring screen.

[0017] Other advantages, objectives and features of the present invention will be set forth in part in the following description and, in part, will be apparent to those skilled in the art based on an examination of the following or may be taught from the practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention, and for ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0019] Figure 1 This is a process operation diagram of a video surveillance system for privacy management of medical and health big data; Figure 2 The figure is a flow chart of a video surveillance system for privacy management of medical and health big data. DETAILED DESCRIPTION

[0020] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the specific implementation mode, structure, characteristics and effects of the present invention are described in detail below in combination with the accompanying drawings and preferred embodiments.

[0021] Embodiment 1 This embodiment describes in detail the specific application of a video surveillance system for privacy management of medical and health big data in monitoring wards and data centers in large comprehensive hospitals. Through the present invention, the hospital can further optimize and customize the system parameters, strategies, etc. according to its own specific layout, workflow, security management requirements and other characteristics to achieve the most ideal monitoring and management effects.

[0022] Multiple cameras with high resolution (such as 4K ultra-high definition), low illumination and integrated advanced scene analysis technology are carefully deployed in key locations in the ward area and data center. After the camera is turned on, the scene analysis technology conducts a comprehensive and detailed real-time analysis of the surrounding environment. For the analysis of the medical environment, the ward area is used as an example to explain in detail. Through the formula To determine whether the area is a ward, a series of representative regional features are selected, such as beds (as the core feature for judging wards), a variety of common medical equipment (such as infusion pumps, ECG monitors, ventilators, etc., the presence of these devices highly indicates the ward area), ward layout (including the specific location of the bathroom, the relative layout of windows and beds, etc.) as regional features (assuming that ), and set weight coefficients based on a large amount of experimental data and in-depth research on ward environmental characteristics, such as bed weight coefficient (for bed characteristics) = 0.3), weight coefficient of main medical equipment (For key equipment such as infusion pumps and ECG monitors to be levied) = 0.25), ward layout weight coefficient (for ward layout features) = 0.2), weight coefficients of other auxiliary features (such as call buttons and oxygen interfaces in the ward) (for other features) = 0.1), etc. By using high-precision image recognition technology to compare the actual scene with the predefined ward area feature template, the matching degree of each feature is accurately calculated. , if the calculated regional matching score If it is higher than the set threshold (such as 0.65), it is highly determined that the camera is located in the ward area. At the same time, the formula Drill down into ward traffic, for example, tracking the number of people who have been in a ward for a specific 10-minute interval Personnel targets, one of which requires medical staff to make frequent ward rounds due to the patient's condition, and the frequency of its occurrence in this time period (for this patient) = 6) times, ward area square meters, considering that the flow of people in the ward is usually regular but occasionally changes due to emergencies, the coefficient of change of flow of people , through real-time statistical analysis of the distribution of personnel in the ward, the personnel distribution variance is obtained (Indicates that the distribution of personnel is relatively uniform but there are certain fluctuations), thus calculating the personnel flow index , in order to accurately judge the activity status of people in the ward, and provide a key reference for the subsequent flexible adjustment of monitoring strategies. For the analysis of lighting conditions, the formula is used and Calculate the light intensity and color temperature separately. In actual operation, in order to obtain more accurate light data, select The light intensity is measured at pixels evenly distributed in the image, and the light intensity value of each pixel is obtained through the photosensitive element inside the camera and a complex light intensity calculation algorithm. , and then calculate the average light intensity , at the same time, select The color temperature of the light is analyzed by using a predefined color temperature conversion coefficient. (Through a large number of experiments and color theory, different color component values ​​have different conversion relationships) to calculate the color temperature of the current light , laying a solid foundation for the precise adjustment of camera parameters.

[0023] The environment and privacy management module carefully formulates targeted monitoring strategies based on the pre-set privacy level information of different areas. Due to the high privacy requirements of the ward area, the monitoring scope is strictly narrowed to the entire ward entrance and public activity area (such as the corridor near the ward entrance), and the image acquisition resolution is greatly reduced (such as from high-definition 1080p to standard definition 720p to reduce the capture of patient privacy details). Through the formula Calculate the monitoring policy index value , assuming that the privacy level parameter of the ward area (Reflecting its extremely high privacy protection needs), data security weight coefficient (relatively focused on privacy protection), privacy protection weight coefficient , calculated Then, the most suitable monitoring strategy for the ward is determined from the pre-stored strategy set. Then, according to the selected monitoring strategy and the real-time lighting conditions, the formula is used to Adjust the camera parameters, assuming the basic camera parameter values ​​based on the ward monitoring strategy Considering that the ward monitoring strategy has a relatively small demand for parameter adjustment, the monitoring strategy influence coefficient Since the lighting changes in the ward are relatively small and the image quality needs to be relatively stable, the lighting conditions affect the coefficient , reference light intensity value (normal lighting during the day is 500 lux) and the current light intensity value accurately perceived in real time by the camera , accurately calculate the adjusted camera parameters to ensure that clear images that meet privacy protection requirements are obtained, which can not only ensure basic security monitoring in the ward, but also minimize the infringement of patient privacy. For the data center area, in view of its extremely high requirements for security, the monitoring range is expanded to the entire data center and its surrounding channels (including entrances, exits, channels between equipment, etc.), significantly improve the image acquisition resolution (such as upgrading to 4K ultra-high definition to clearly capture details such as equipment status and personnel operations), and fully enable the intelligent recognition function (such as accurate recognition of personnel identity, behavior, equipment operation status, etc.). The strategy is also determined by the above monitoring strategy selection formula, and then the camera parameters are adjusted. For example, in order to obtain clearer and brighter images to meet the strict monitoring requirements of the data center, the aperture is increased, the sensitivity is increased (such as adjusted to ISO800), etc. At the same time, according to the lighting characteristics of different areas inside the data center (such as the server cabinet area is relatively dark, and the channel area is relatively bright) and the monitoring focus (such as the monitoring of the operation status of server equipment requires higher accuracy), the camera parameters are further optimized to ensure that all key activities in the data center can be effectively monitored and any possible security threats can be discovered and responded to in a timely manner.

[0024] The data monitoring module efficiently receives and processes the video image data from the environment and privacy management module. During the data transmission process, it uses reliable and efficient communication protocols (such as TCP / IP protocol, and uses encrypted transmission channels to ensure data security) and powerful data verification mechanisms (such as CRC verification combined with data retransmission mechanism to ensure data integrity and accuracy) to ensure stable and complete data transmission, and prevent data loss, damage or tampering. Convert the data into a format suitable for healthcare big data analysis, assuming the original data format is A common video encoding format (such as MJPEG) that is not suitable for big data analysis. The specific data specifications required for medical and health big data analysis The corresponding data specifications in the original data format are the specified image resolution (such as 1920×1080 pixels), frame rate (such as 30fps), encoding format (such as H.264 high-efficiency encoding format), etc. (For example, the original resolution of the camera may be 1280×720 pixels, the frame rate may be 25fps, and the encoding format may be MJPEG). The data format adjustment factor is set according to the degree of data format difference and the urgency and importance of the conversion. (Indicates that the response to format conversion is more sensitive and needs to be adjusted quickly and significantly), a key data indicator in the original data format (such as original resolution ratio, frame rate ratio, etc.) and the corresponding key data indicators required for medical and health big data analysis , accurately calculate the converted video image data format , so that it fully meets the requirements of the medical and health big data analysis platform, facilitates subsequent efficient data analysis, mining and utilization, and uses advanced encryption algorithms Data involving patient privacy (such as patient facial images, detailed medical records, genetic testing data and other highly sensitive information) is highly encrypted. Assume that the generated symmetric encryption key It is a 128-bit or 256-bit binary sequence (select the appropriate key length according to the sensitivity of the data), the key obfuscation coefficient (Adding extra security and complexity to data encryption) by calculating the key through advanced hash functions such as SHA-256 or higher hash algorithms Hash value , for plaintext data Encrypt to get ciphertext , ensuring the security of data during storage, transmission and analysis, effectively preventing patients' privacy data from being accessed, stolen or modified without authorization, and providing solid protection for patient information security.

[0025] Facial recognition module carefully installs high-precision facial recognition equipment at the entrance of the hospital data management area and key locations around it to ensure that the facial features of people entering can be accurately captured. Before the system is activated, the facial feature data of legal management personnel (including doctors, nurses, system administrators, medical data analysts and other personnel with permission to access the data center) are fully collected and securely stored as benchmark data. The collection process uses high-resolution and multi-angle image acquisition equipment to ensure the comprehensiveness and accuracy of facial features, including subtle features of facial contours, precise shapes and positions of facial features, facial texture and other detailed information. When a person enters the area, a high-precision image acquisition device is used to collect their facial images in real time, and the formula is used to calculate the facial features of the person. The real-time image is carefully compared with the benchmark data. Assuming that a rich set of key facial feature points such as eyes, nose, mouth, eyebrows, and facial contour key points are selected ( ), and set different weight coefficients according to the importance and uniqueness of these feature points in facial recognition (such as the weight coefficient of the eye feature point (for eye feature points) = 0.3), weight coefficient for nose feature points (for the nose feature point) = 0.25), etc.), taking into full account the impact of natural movements that may occur when a person enters (such as slight shaking during normal walking, turning the head, etc.) on the recognition results, the action influence coefficient , calculate the facial matching score ,like Higher than or equal to the set threshold If the time is lower than the threshold, the person is considered to be an unauthorized person and a strong alarm is immediately issued (such as connecting to the hospital security system, triggering a high-decibel sound and light alarm, and displaying the alarm location and facial image of the person in the security monitoring center). At the same time, the relevant information is recorded in detail (such as entry time, facial image, number of attempts to enter, etc.) to facilitate subsequent investigation and processing, effectively preventing unauthorized persons from entering the data center and ensuring the security of medical data.

[0026] When the privacy display control module monitors the screen and transmits it to the monitoring display terminal, it starts working immediately. Accurately identify the privacy areas in the monitoring images. Taking the ward images as an example, we select the patient's bed (especially the area covered by the patient's body), the equipment and operation area where diagnosis and treatment are being performed (such as the area where surgery or examination is being performed), and the patient's private areas (such as when the face is blocked or blurred) as key image content elements. ), and set the weight coefficient according to its close relationship with patient privacy (such as patient bed weight coefficient (for patient beds) = 0.4), weight coefficient of diagnosis and treatment equipment and operation area (for diagnostic equipment and operation areas) = ​​0.3), etc.), through in-depth analysis and statistics of a large number of ward monitoring images, the probability of each element being present is obtained. , regional area influence coefficient (This means that larger privacy-related areas are easier to identify) and accurately calculate the privacy area identification index. ,like If the value is higher than the set threshold (such as 0.65), the area is determined to be a privacy area. According to the pre-stored permission information corresponding to different personnel roles, the formula Determine the display degree of the privacy area, assuming the doctor role permission parameters (for doctors) = 1 (indicates that doctors need to fully understand the patient's condition during diagnosis and treatment and can view the complete privacy area), nurse role permission parameters (For nurses) = 0.8 (indicates that nurses can view some private areas when performing nursing tasks, but there may be a certain degree of blur or occlusion to protect patient privacy while meeting nursing needs), logistics staff role permission parameters (For logistics personnel) = 0 (indicates that logistics personnel do not need to view the private area, their work is mainly focused on public areas and facility maintenance, and does not involve patient privacy information), the initial occlusion degree of the private area (For example, the blur level is 0.5 or mosaic masking is used, which can be adjusted according to specific privacy requirements and display effects), permission adjustment coefficient (Indicates that the display degree of the privacy area is adjusted significantly according to the difference in personnel role permissions), default role code , according to the role coding of different personnel when viewing the monitoring screen , accurately calculate the final display degree of the privacy area , achieving highly differentiated privacy management. For example, when a doctor views it, the privacy area is fully displayed so that they can fully understand the patient's condition and treatment. When a nurse views it, the privacy area is partially displayed (such as the blur is reduced or the mosaic occlusion ratio is reduced), which can both assist nursing work and protect patient privacy. When logistics staff view it, the privacy area is completely blocked to ensure that they cannot obtain any patient privacy information.

[0027] To sum up, through the present invention, the hospital can further optimize and customize the system parameters, strategies, etc. according to its own specific layout, workflow, security management requirements and other characteristics, so as to achieve the most ideal monitoring and management effects, ensure the safe and efficient operation of the hospital and the strict confidentiality of patient information.

[0028] Embodiment 2 This embodiment describes in detail the application of a video surveillance system for privacy management of medical and health big data in monitoring consulting rooms and drug storage rooms in small specialized hospitals. Through the present invention, patient privacy and drug management security are effectively protected, providing strong support for medical services and management.

[0029] The image acquisition module arranges cameras in the treatment room and the drug storage room, and the cameras have high resolution, low illumination and field analysis capabilities. In terms of medical environment analysis, for the treatment room, the formula is used Determine the regional type, select the diagnosis and treatment bed, examination instruments (such as stethoscopes, sphygmomanometers, etc.), medical waste bins, etc. as regional features, and set the diagnosis and treatment bed weight coefficient (for diagnosis and treatment bed symbol) = 0.4), check the instrument weight coefficient (for inspection instrument characteristics) = 0.35), weight coefficient of medical waste barrel (for medical waste barrel feature) = 0.15), other feature weight coefficients (for other symbols) = 0.1), calculate feature matching degree through image recognition , determine whether the area where the camera is located is a treatment room, and at the same time, through the formula Analyze the flow of personnel in the diagnosis and treatment room to provide a basis for adjusting the monitoring strategy. For the drug storage room, select shelves, drug refrigeration equipment, fire protection facilities, etc. as regional features, set corresponding weight coefficients, determine its regional type, and analyze the flow of personnel (usually there is less flow of personnel). In the analysis of lighting conditions, the same and Calculate light intensity and color temperature to prepare for camera parameter adjustment.

[0030] The environment and privacy management module formulates corresponding monitoring strategies for the high privacy level of the clinic, limits the monitoring range to the door of the clinic and the surrounding public areas, reduces the image acquisition resolution (such as 960×540 pixels), and uses the formula Calculate the monitoring policy index value , assuming that the privacy level parameter of the clinic is , data security weight coefficient , privacy protection weight coefficient , then according to the formula Adjust the camera parameters, assuming the basic camera parameter values (If the initial aperture value is , initial sensitivity value is ISO150, etc.), monitoring strategy influence coefficient , Lighting condition influence coefficient (The lighting in the clinic is relatively stable), reference light intensity value (Normal lighting during the day is 400 lux) and the current light intensity value , calculate the adjusted parameters, focus on safety monitoring in the drug storage room, expand the monitoring range to the entire room, improve the image acquisition resolution (such as increasing it to 2K resolution), turn on the intelligent recognition function, determine the monitoring strategy and adjust the camera parameters through the above formula, such as increasing the aperture, increasing the sensitivity (such as adjusting to ISO600), etc., to ensure that the drug storage situation and personnel activities can be clearly monitored.

[0031] The data monitoring module receives and processes video image data, uses communication protocols (such as UDP protocol) and data verification mechanisms (such as parity check) to ensure data transmission, and uses the formula Convert data format, assuming the original data cell , specific data specifications required for medical and health big data analysis (such as resolution, frame rate, etc.), the corresponding data specifications in the original data format , data format adjustment coefficient (Set according to data differences and requirements), key data indicators in the original data format And the corresponding key data indicators required , calculate the converted format , using encryption algorithm Encryption of data involving patient privacy (such as medical records) and drug information (such as drug name, quantity, expiration date, etc.), assuming symmetric encryption key is 256 bits, the key obfuscation coefficient , calculate the key hash value , for plaintext data Encrypted to get ciphertext .

[0032] Facial recognition module sets up facial recognition equipment at the entrance of the drug storage room to collect and store the facial feature data of legal management personnel (such as pharmacy staff, drug administrators, etc.) as benchmark data. When a person enters, facial images are collected in real time and the formula is used to identify the person. For comparison, assume that key facial feature points are selected and weight coefficients are set (such as eye feature point weight coefficient (for eye feature points) = 0.32), weight coefficient for nose feature points (for the nose feature point) = 0.28), etc.), motion influence coefficient (Considering that the movement of a person entering is relatively small), calculate the facial matching score ,like If the value is higher than or equal to the set threshold (such as 0.2), the person is determined to be an authorized person and is allowed to enter. If the value is lower than the threshold, the person is determined to be an unauthorized person and an alarm is issued (such as connecting to a hospital security system, triggering an alarm to notify security personnel and recording relevant information (such as entry time, facial image, and related circumstances of attempted entry).

[0033] For the privacy display control module, when the monitoring screen is transmitted to the monitoring display terminal, the privacy display control module starts to work. For the monitoring screen of the treatment room, the formula Identify privacy areas, select patient treatment areas, medical records placement areas, etc. as image content elements, and set weight coefficients (such as the weight coefficient of the patient's treatment site (for the patient's diagnosis and treatment area) = 0.45), weight coefficient of the medical record placement area (for the medical records area) = 0.35), etc.), by analyzing a large number of monitoring images of the treatment rooms, the probability of each element being present is counted , regional area influence coefficient , calculate the privacy region identification index ,like If the threshold is higher than 0.6, it is considered as a privacy area. According to the pre-stored permission information corresponding to different personnel roles, the formula Determine the degree of privacy area display, assuming the doctor role permission parameters (for doctors) = 1), nurse role permission parameters (for nurses) = 0.7) (indicates that some private areas can be viewed, but the degree of blur is high), other non-medical personnel role permission parameters (for other people) = 0), the initial occlusion degree of the privacy area (For example, the degree of fuzziness is 0.6), the authority adjustment coefficient , default role encoding , according to the role coding of different personnel when viewing the monitoring screen , calculate the final display degree of the privacy area For example, when a doctor views the privacy area, it is fully displayed; when a nurse views the privacy area, it is partially displayed and has a high degree of blur; when other non-medical personnel view the privacy area, it is completely blocked. For the monitoring screen of the drug storage room, special drugs (such as storage areas for narcotic drugs and psychotropic drugs) or key operation areas involving drug management (such as drug entry and exit registration areas) are mainly identified as privacy areas, and corresponding image content elements and weight coefficients are set for privacy area identification. Permission parameters are set according to different roles such as drug management personnel, hospital logistics support personnel (responsible for non-critical management work such as drug handling), and other unrelated personnel. For example, the role permission parameters of drug management personnel (For drug management personnel) = 1), role permission parameters for logistics support personnel (For logistics support personnel) = 0.5) (indicates that some areas can be viewed, but the key drug storage area is highly obscured), other irrelevant personnel role permission parameters (for other irrelevant personnel) = 0), the display degree of the privacy area is calculated by the above formula to realize the privacy management of the monitoring screen of the drug storage room, ensuring the security of drug management information while meeting the reasonable viewing needs of different personnel in the drug management process.

[0034] In summary, the application of the present invention in small specialized hospitals has fully demonstrated its ability to adapt to the functional requirements of different regions, protect privacy, ensure security and support big data analysis, effectively improving the management level and security of the hospital, and providing strong technical support for the daily operation of the hospital.

[0035] The above description is only a preferred embodiment of the present invention and does not limit the present invention in any form. Although the present invention has been disclosed as a preferred embodiment as above, it is not used to limit the present invention. Any technical personnel in this field can make some changes or modify the technical contents disclosed above into equivalent embodiments without departing from the scope of the technical solution of the present invention. However, any brief modifications, equivalent changes and modifications made to the above embodiments based on the technical essence of the present invention without departing from the content of the technical solution of the present invention are still within the scope of the technical solution of the present invention.

Claims

1. A video surveillance system for privacy management of medical health big data, characterized in that: The system includes the following components: image acquisition module, environment and privacy management module, data monitoring module, face recognition module and privacy display control module; The image acquisition module is composed of multiple cameras distributed in various areas of the medical site. The cameras have high resolution and low illumination characteristics, and have scene analysis capabilities. They can automatically analyze the medical environment and lighting conditions in real time, and transmit the analysis results to subsequent modules, providing the original video image data source for the entire monitoring system; The environment and privacy management module formulates monitoring strategies for each area based on the privacy level information of different areas. For areas with high privacy requirements, the monitoring scope is limited to the ward entrance and public activity area to avoid excessive monitoring of the patient's private rest area in the ward. At the same time, the resolution of image acquisition is reduced so that the collected images meet the security monitoring requirements while not causing the patient's privacy details to be exposed. For key areas, considering their impact on medical data security and system operation stability, the area is monitored in all directions and the resolution of image acquisition is improved. At the same time, the intelligent recognition function is turned on to provide real-time warnings for abnormal personnel and behaviors through image recognition technology. Once unauthorized entry or abnormal behavior is found, an alarm is issued in time and corresponding measures are taken. For monitoring strategies, the camera parameters are finely adjusted according to the pre-constructed lighting conditions and parameter mapping relationship table. The adjusted camera parameters cover multiple aspects, including but not limited to focal length, aperture, sensitivity and white balance. The data monitoring module is responsible for receiving and processing video image data after environmental and privacy management, using communication protocols and data verification mechanisms to ensure stable and complete data transmission, and converting data into a format suitable for medical and health big data analysis through algorithms. At the same time, encryption algorithms are used to encrypt the parts involving patient privacy, and the processed video images are displayed in real time through the multi-screen split display function, and specific screens can be enlarged, reduced, and paused according to the operating instructions of the monitoring personnel; The facial recognition module is set at the entrance of the medical data management area and key locations around it. Before the system is activated, facial feature data of legal management personnel are collected and stored as reference data. When a person is detected entering the area, a high-precision image acquisition device is used to collect their facial images in real time. The real-time image is compared with the reference data through a facial recognition algorithm, and the factors of various facial feature points are comprehensively considered. If the comparison is successful, it is determined that the person is authorized to enter. If it fails, it is determined that the person is unauthorized and an alarm is issued while recording relevant information; The privacy display control module uses image recognition technology to identify the content involving privacy areas in the monitoring screen, and uses image processing algorithms to dynamically block it, making part of the content invisible. At the same time, based on the pre-stored permission information corresponding to different personnel roles, when the monitoring personnel views the monitoring screen, they determine whether to display and the degree of display according to their role permissions, thereby achieving differentiated privacy management.

2. A video surveillance system for privacy management of medical health big data according to claim 1, characterized in that: The image acquisition module automatically performs real-time analysis of the medical environment and lighting conditions. For the medical environment analysis, the formula Determine whether the area is a specific medical area, where: Represents the area matching score, which is used to measure the matching degree between the current camera area and various predefined medical areas. is the number of features used for region recognition, It is The weight coefficient of regional features, It is The matching degree of regional features is calculated using the formula Analyze the personnel turnover, including: It represents the personnel flow index, which is used to reflect the density and changes of personnel flow in the current region. is the number of human targets tracked within a certain time interval, It is The frequency with which a person target appears in a specific time period. is the area covered by the current camera monitoring range. is the coefficient of change of personnel turnover, is the variance of personnel distribution. For lighting conditions, the formula is used Analyze the light intensity, where Indicates the average light intensity currently perceived by the camera. is the number of image pixels selected to measure light intensity, It is The light intensity value of the selected pixel point is calculated using the formula Analyze the light color temperature, where Indicates the color temperature of the current light. is the number of image pixels selected for analyzing the light color temperature, It is The color component value of the selected pixel in the RGB color space, It is the color temperature conversion coefficient corresponding to the RGB color component value.

3. A video surveillance system for privacy management of medical health big data according to claim 1, characterized in that: The environment and privacy management module formulates a monitoring strategy for each area based on the privacy level information of different areas. The algorithm formula is: ,in, Indicates the index value of the monitoring strategy finally selected, corresponding to a specific strategy in the set of pre-stored different monitoring strategies. is the privacy level parameter, is the data security weight coefficient, is the privacy protection weight coefficient.

4. A video surveillance system for privacy management of medical health big data according to claim 1, characterized in that: The environment and privacy management module fine-tunes the camera parameters for the monitoring strategy based on the pre-built lighting condition and parameter mapping relationship table. The calculation formula is: ,in, is the adjusted camera parameter value, including focal length, aperture, sensitivity and white balance, It is the basic camera parameter value based on the current regional monitoring strategy. is the monitoring strategy influence coefficient, is the influence coefficient of lighting conditions, is the monitoring strategy index value obtained by the monitoring strategy selection formula, is the average value of all monitoring strategy index values, and They are the reference light intensity value and the current light intensity value respectively.

5. A video surveillance system for privacy management of medical health big data according to claim 1, characterized in that: The data monitoring module converts the data into a format suitable for medical and health big data analysis through an algorithm. The algorithm formula is: ,in, Indicates the converted video image data format. It is the original video image data format, that is, the unconverted data format received from the environment and privacy management module. It is a specific data specification required for medical and health big data analysis. is the corresponding data specification in the original data format, is the data format adjustment factor, is a key data indicator in the original data format, It is the corresponding key data indicator required for medical and health big data analysis.

6. A video surveillance system for privacy management of medical health big data according to claim 1, characterized in that: The data monitoring module uses an encryption algorithm to encrypt the part involving patient privacy, and the algorithm formula of the encryption algorithm is: ,in, Indicates the encrypted ciphertext, that is, the data after encryption. Indicates the key For plain text To perform encryption operations, Is the plaintext data that needs to be encrypted. is the symmetric encryption key, is the key obfuscation coefficient, Is the key The hash value of .

7. A video surveillance system for privacy management of medical health big data according to claim 1, characterized in that: The facial recognition module uses a high-precision image acquisition device to collect facial images in real time, and compares the real-time image with the reference data through a facial recognition algorithm, the formula of which is: ,in, Represents the facial matching score, which is an indicator that comprehensively evaluates the matching degree between the real-time acquired facial image and the pre-stored benchmark data. Less than threshold , then it is judged as a comparison failure, that is, it is considered to be an unauthorized person. If Greater than or equal to , then the comparison is judged to be successful, that is, it is considered to be an authorized person, It is the first facial image collected in real time. The eigenvalues ​​of feature points, It is the first one in the pre-stored benchmark data. The eigenvalues ​​of feature points, It is The weight coefficient of feature points, is the total number of selected facial feature points, is the action influence coefficient, It is The variation range of a feature point under the natural movement of the person.

8. A video surveillance system for privacy management of medical health big data according to claim 1, characterized in that: The privacy display control module uses image recognition technology to identify the content involving the privacy area in the monitoring screen. Specifically, the formula Accurately identify the privacy area in the monitoring screen, including: It represents the privacy area identification index, which is an indicator for comprehensively evaluating whether a certain area in the monitoring image is a privacy area. Above the set threshold , then the area is determined to be a privacy area. If Below threshold , then the area is determined not to be a privacy area. It is The weight coefficients of the image content elements, It is the first The existence probability of the image content elements, is the total number of image content elements selected, is the area influence coefficient, It is The area ratio of the area where the image content element is located is the ratio of the area where the image content element is located to the area of ​​the entire monitoring screen.

9. A video surveillance system for privacy management of medical health big data according to claim 1, characterized in that: The privacy display control module determines whether to display and the degree of display according to the pre-stored authority information corresponding to different personnel roles when the monitoring personnel view the monitoring screen according to their role authority. The determination formula for the degree of display is: ,in, Indicates the final display degree of the privacy area. It is the personnel role permission parameter. is the initial occlusion degree of the privacy area, is the authority adjustment coefficient, It is the role code of the person currently viewing the surveillance screen. The code can be used to quickly identify the person's role type. is the default role encoding.

Images