Patient management system and method for pneumology department based on video monitoring
Through the processing of video frame pixel matrix and multiple algorithms, the problem of inaccurate respiratory detection in video surveillance is solved, and accurate monitoring of respiratory frequency and mode is achieved, false alarms and missed reports are reduced, and the respiratory patient management system is improved. Accuracy and personalized intervention capabilities.
Patent Information
- Application Number
- CN202411505251.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-10-28
- Publication Date
- 2024-11-26
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
In video surveillance, changes in non-respiratory-related movements and respiratory frequency of patients lead to inaccurate detection results, and there are individual differences in respiratory patterns of different patients, which can easily lead to false alarms.
By collecting video frame pixel matrix in real time, using breathing area detection and extraction algorithms, adaptive recursive smoothing algorithms, segmented hybrid transformation algorithms and personalized breathing abnormality detection algorithms, calculate breathing activity indicators and frequency, dynamically adjust detection thresholds, and achieve accurate Respiratory status monitoring and abnormal detection.
It improves the accuracy and robustness of respiratory detection, reduces false alarms and missed reports, realizes accurate monitoring and personalized intervention in respiratory patterns of different patients, and improves the timeliness and reliability of medical management.
Smart Images

Figure CN119028575A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of medical management, and in particular to a respiratory patient management system and method based on video monitoring. Background Art
[0002] Respiratory diseases are a common public health problem worldwide. Especially today, with the incidence of diseases such as chronic obstructive pulmonary disease, asthma and pneumonia increasing year by year, medical institutions need effective management methods for respiratory patients, not only to monitor the patient's condition in a timely manner, but also to ensure that emergency measures can be provided quickly in the event of an acute attack. Traditional medical monitoring methods rely on medical staff to patrol the ward, which is easily limited by human resources and cannot achieve real-time monitoring around the clock; at the same time, the use of mechanical monitoring equipment requires active cooperation from patients, and the ability of continuous and non-contact monitoring is weak. Against the background of the rapid development of information technology and artificial intelligence, the respiratory patient management system based on video monitoring has gradually become an emerging solution that can monitor the patient's physiological state and behavior patterns through video data, thereby improving the accuracy, intelligence and timeliness of respiratory disease patient management.
[0003] With the continuous advancement of technology, the respiratory patient management system based on video surveillance will become an important part of future medical management, bringing profound changes to medical management.
[0004] However, the above technology has the following technical problems: in video surveillance, the patient's body may produce non-respiration-related movements, such as arm swinging and head turning, which may lead to inaccurate respiratory detection results; the respiratory rate varies greatly in different states, such as being relatively stable when resting, but may become faster or irregular during activity or pathological conditions, and short-term and violent fluctuations in the respiratory signal may easily lead to false alarms, such as occasional coughing or snoring; there are significant individual differences in the breathing patterns of different patients, and the respiratory rate is affected by the circadian rhythm, which may easily lead to false alarms. Summary of the invention
[0005] The present invention provides a respiratory patient management system and method based on video surveillance to solve the following technical problems: in video surveillance, the patient's body may produce non-respiration-related movements, such as arm swinging and head turning, which may lead to inaccurate respiratory detection results; the respiratory rate varies greatly in different states, such as being relatively stable when resting, but may become faster or irregular during activities or pathological states, and short-term and violent fluctuations in respiratory signals may easily lead to false alarms, such as occasional coughing or snoring; there are significant individual differences in the breathing patterns of different patients, and the respiratory rate is affected by the circadian rhythm, which may easily lead to false alarms.
[0006] A respiratory patient management system and method based on video surveillance of the present invention specifically includes the following technical solutions: A method for managing respiratory patients based on video monitoring comprises the following steps: S1. Real-time acquisition of patient images to obtain a video frame pixel matrix; based on the video frame pixel matrix, the pixel change amount and the respiratory region matrix are calculated by a respiratory region detection and extraction algorithm; S2, based on the pixel change and the respiratory area matrix, the respiratory activity index is calculated by an adaptive recursive smoothing algorithm; S3, based on the respiratory activity index, the respiratory frequency is calculated by a segmented hybrid transformation algorithm; S4. Calculate a personalized detection threshold through a personalized breathing abnormality detection algorithm; judge the breathing frequency based on the personalized detection threshold to obtain a judgment result; and adopt an intervention strategy based on the judgment result.
[0007] Preferably, the S1 specifically includes: In the process of implementing the breathing area detection and extraction algorithm, nonlinear compression and Gaussian weight enhancement are introduced to obtain the pixel change amount. The specific calculation formula of the pixel change amount is: , in, Indicates at location The pixel change at ; represents the hyperbolic tangent function used to compress the magnitude of pixel changes; Indicates at location Gaussian weight at ; Indicates the height of the video frame; Indicates the width of the video frame; represents the standard deviation of the Gaussian function; express Always at the location The pixel value at ; express Always at the location The pixel value at ; A threshold screening mechanism is introduced to obtain the breathing region matrix.
[0008] Preferably, the S2 specifically includes: The adaptive recursive smoothing algorithm introduces an adaptive smoothing coefficient and a recursive memory coefficient to calculate the respiratory activity index; the specific calculation formula is: , in, Indicated in Respiratory activity indicators at all times; represents the adaptive smoothing coefficient; represents the breathing zone matrix; Indicates at location The amount of pixel change at ; Indicated in Respiratory activity indicators at all times; Represents the recursive memory coefficient.
[0009] Preferably, the S2 specifically includes: The sensitivity coefficient is introduced, and the adaptive smoothing coefficient is adjusted in combination with the respiratory activity index at the previous moment.
[0010] Preferably, the S2 specifically includes: The weight coefficient is introduced to scale the respiratory activity index at the previous moment to obtain the recursive memory coefficient.
[0011] Preferably, the S3 specifically includes: In the process of implementing the segmented hybrid transformation algorithm, the segmented transformation function is introduced to dynamically adapt to the changes in the patient's respiratory state, calculate the rate of change of the respiratory activity index over time, and obtain the respiratory frequency; the specific formula of the segmented transformation function is: , in, Indicated in Piecewise transformation function of respiratory activity at each moment; and denote the first weight parameter and the second weight parameter, respectively, corresponding to different breathing states; and represent a first amplitude coefficient and a second amplitude coefficient, respectively, and are used to adjust the amplitudes of the sine function and the cosine function, respectively; and represent the first frequency parameter and the second frequency parameter, respectively, and are used to control the periods of the sine function and the cosine function, respectively; Indicates the threshold of respiratory activity strength.
[0012] Preferably, the S4 specifically includes: In the process of implementing the personalized breathing anomaly detection algorithm, physiological rhythm modulation items and adaptive adjustment items are introduced to obtain personalized detection thresholds. Based on the personalized detection thresholds, the breathing pattern is monitored and anomaly detected in real time. The calculation formula of the personalized detection threshold is: , in, Indicated in Personalized detection threshold at each moment; represents the basic detection threshold; represents the circadian rhythm modulation term; represents the modulation coefficient of the physiological rhythm; Indicates the length of the day-night cycle; Represents an adaptive adjustment item; Indicates the historical average respiratory activity index; Represents the sensitivity parameter.
[0013] A respiratory patient management system based on video surveillance, including the following parts: Video acquisition module, respiratory area detection and extraction module, respiratory activity index calculation module, respiratory rate calculation module, personalized abnormality detection and intervention module; The video acquisition module acquires patient images in real time to obtain a video frame pixel matrix; the video frame pixel matrix is output to the respiratory area detection and extraction module; The breathing area detection and extraction module calculates the pixel change amount and the breathing area matrix based on the video frame pixel matrix through the breathing area detection and extraction algorithm; and outputs the pixel change amount and the breathing area matrix to the breathing activity index calculation module; The respiratory activity index calculation module calculates the respiratory activity index based on the pixel change amount and the respiratory area matrix through an adaptive recursive smoothing algorithm; and outputs the respiratory activity index to the respiratory frequency calculation module; The respiratory rate calculation module calculates the respiratory rate based on the respiratory activity index through the segmented hybrid transformation algorithm; the respiratory rate is output to the personalized abnormality detection and intervention module; The personalized abnormality detection and intervention module calculates a personalized detection threshold based on the respiratory frequency through a personalized respiratory abnormality detection algorithm; based on the personalized detection threshold, the current respiratory frequency is judged to obtain a judgment result, and an intervention strategy is adopted according to the judgment result.
[0014] The beneficial effects of the technical solution of the present invention are: 1. Construct a video frame pixel matrix, and achieve accurate identification of breathing-related areas through respiratory area detection and extraction algorithm. By analyzing the video frame pixel matrix, nonlinear compression and Gaussian weight enhancement are used to accurately extract the respiratory area, reducing interference from irrelevant areas such as arms and head, ensuring the focus of the detection process, and being able to capture the motion characteristics of core areas such as the chest, avoiding false alarms caused by small movements and environmental noise, and improving the accuracy and robustness of detection.
[0015] 2. The adaptive recursive smoothing algorithm can achieve accurate tracking and fluctuation smoothing of respiratory activities. Combined with the adaptive smoothing coefficient and the recursive memory coefficient, it can not only capture sudden respiratory changes in the short term, but also retain the historical state, reduce misjudgment caused by accidental fluctuations, and enable the respiratory patient management system based on video surveillance to respond quickly in emergency situations. At the same time, it can avoid misjudgment caused by temporary interference factors (such as slight body movement or equipment shaking), providing high stability and flexibility for respiratory status assessment.
[0016] 3. The personalized respiratory frequency under different breathing states is calculated through the segmented hybrid transformation algorithm. According to the changes in respiratory activity indicators, different algorithms are dynamically switched to handle steady breathing and complex breathing states (such as rapid breathing). The frequency calculation is accurately adapted through quadratic functions, sinusoidal functions, logarithmic functions, etc., to adapt to the real-time changes in the patient's breathing, and can accurately reflect the natural breathing cycle or sudden breathing abnormalities. It is especially suitable for the different breathing patterns of patients during sleep and activity, ensuring personalized and dynamic adaptation capabilities.
[0017] 4. Through the personalized abnormal breathing detection algorithm, real-time adjustment of personalized detection thresholds is achieved. According to the patient's historical respiratory activity and circadian rhythm, the personalized detection threshold is dynamically adjusted through physiological rhythm modulation items and adaptive update mechanisms to make the detection results fit the individual's breathing pattern, avoiding missed reports and false alarms caused by a single threshold. It can accurately judge the abnormal breathing state and provide medical staff with real-time and reliable test results, thereby improving the accuracy and response efficiency of abnormal detection.
[0018] 5. Through intelligent intervention strategies, timely intervention and risk management are achieved. The personalized detection threshold is compared with the difference between the current respiratory rate and the historical average respiratory rate to obtain the judgment result. According to the judgment result, the intervention strategy is automatically generated. When abnormal breathing is detected, an alarm is immediately triggered and emergency intervention is triggered. Otherwise, regular observation is performed to ensure that medical staff can take timely response measures to avoid the potential risks caused by the failure to detect abnormal breathing in time, thereby improving the reliability of the respiratory patient management system based on video surveillance in patient safety management. BRIEF DESCRIPTION OF THE DRAWINGS
[0019] Figure 1 This is a structural diagram of a respiratory patient management system based on video monitoring according to the present invention; Figure 2 This is a flow chart of a respiratory patient management method based on video monitoring described in the present invention. 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 technical scheme in the embodiment of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiment of the present invention. Obviously, the described embodiment is only a part of the embodiment of the present invention, not all 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.
[0021] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.
[0022] The specific scheme of a respiratory patient management system and method based on video monitoring provided by the present invention is described in detail below with reference to the accompanying drawings.
[0023] See attached Figure 1 , which shows a structure diagram of a respiratory patient management system based on video monitoring provided by an embodiment of the present invention, the system includes the following parts: Video acquisition module, respiratory area detection and extraction module, respiratory activity index calculation module, respiratory rate calculation module, personalized abnormality detection and intervention module; The video acquisition module acquires patient images in real time through video monitoring equipment to obtain a video frame pixel matrix; the video frame pixel matrix is output to the respiratory area detection and extraction module; The breathing area detection and extraction module calculates the pixel change amount and the breathing area matrix based on the video frame pixel matrix through the breathing area detection and extraction algorithm; and outputs the pixel change amount and the breathing area matrix to the breathing activity index calculation module; The respiratory activity index calculation module calculates the respiratory activity index based on the pixel change amount and the respiratory area matrix through an adaptive recursive smoothing algorithm; and outputs the respiratory activity index to the respiratory frequency calculation module; The respiratory rate calculation module calculates the respiratory rate based on the respiratory activity index through the segmented hybrid transformation algorithm; the respiratory rate is output to the personalized abnormality detection and intervention module; The personalized abnormality detection and intervention module calculates a personalized detection threshold based on the respiratory frequency through a personalized respiratory abnormality detection algorithm; based on the personalized detection threshold, the current respiratory frequency is judged to obtain a judgment result, and an intervention strategy is adopted according to the judgment result.
[0024] See attached Figure 2 , which shows a flow chart of a respiratory patient management method based on video monitoring provided by an embodiment of the present invention, the method comprising the following steps: S1. Real-time acquisition of patient images to obtain a video frame pixel matrix; based on the video frame pixel matrix, the pixel change amount and the respiratory region matrix are calculated by a respiratory region detection and extraction algorithm; The patient's image is collected in real time by video monitoring equipment, and the video frame pixel matrix is obtained, which is expressed as follows: , in, Indicated in The video frame pixel matrix at time; Indicates the position in the video frame pixel matrix is Pixels at ; The RGB (red, green, and blue) color space represents pixel values, where each dimension ranges from 0 to 255; Indicates the vertical (height) range of the video frame pixels. Indicates the height of the video frame (number of pixels), which is determined by the resolution of the video surveillance device; Indicates the horizontal (width) range of the video frame pixels. Indicates the width of the video frame (number of pixels), which is determined by the resolution of the video surveillance device; Based on the video frame pixel matrix, the pixel change and the breathing area matrix are calculated by the breathing area detection and extraction algorithm; The pixels of each frame are compared with the pixels at the corresponding position of the previous frame, and the differences are calculated to detect the changes in pixel values between adjacent video frames, that is, the pixel change amount, and find the breathing-related areas; in order to further reduce interference and noise, a two-layer enhancement and filtering strategy is adopted: first, nonlinear compression is used to compress the values of pixel changes in adjacent video frames through a hyperbolic tangent function so that small changes will not cause false alarms; second, the influence of pixels in the central area is enhanced through Gaussian weights, and the central part of the video screen is given a higher weight, while the weight of the edge part will gradually decrease, ensuring that movements related to the core parts of the patient's body (such as the chest) are detected first, while movements related to interference factors at the edge of the body (such as movement of the arms or head) are suppressed; The calculation formula for pixel change is: , in, Indicates at location The pixel change at reflects the degree of change of pixel values in two adjacent video frames; Represents the hyperbolic tangent function, which is used to compress the amplitude of pixel changes, smooth small changes, avoid noise effects, and highlight significant changes; Indicates at location The Gaussian weight at is used to enhance the influence of the central pixels in the breathing area and reduce the interference in the edge area; and Respectively represent the center position of the height and width of the video frame, which are used to locate the geometric center of the video frame to enhance the accuracy of central area detection; Represents the standard deviation of the Gaussian function, which is used to control the distribution range of the Gaussian weight. It can be set according to the specific implementation scenario and is not limited here; express At the time (current frame) at position The pixel value at ; express At the time (previous frame) at position The pixel value at ; Indicates that the current frame is at the same position as the previous frame. The absolute difference of the pixel values at is used to detect the change of pixel values in two adjacent video frames and identify breathing-related motion; The threshold screening mechanism is introduced to exclude irrelevant small movements by setting a reasonable breathing change threshold, and obtain the breathing area matrix; the breathing area matrix contains all pixels whose changes exceed the breathing change threshold. The location ensures that the test results of the respiratory patient management system based on video surveillance are accurate and reliable; , in, Represents the breathing area matrix, which contains all pixels whose changes exceed the breathing change threshold location; Respiratory change threshold, used to exclude minor non-respiratory related movements, is determined according to the standards set by medical staff and is not limited here; The above steps can effectively deal with various interference factors in respiratory monitoring and provide reliable data support for subsequent management; S2, based on the pixel change and the respiratory area matrix, the respiratory activity index is calculated by an adaptive recursive smoothing algorithm; Based on the pixel variation and the respiratory region matrix, the respiratory activity index is obtained through an adaptive recursive smoothing algorithm; The adaptive recursive smoothing algorithm adopts a recursive calculation method, combining an adaptive smoothing coefficient and a recursive memory coefficient. The adaptive smoothing coefficient is used to determine the importance of the current video frame data, while the recursive memory coefficient is used to ensure that the historical state is retained in the calculation, so that short-term and long-term changes can be organically combined. Specifically, the respiratory activity index at each moment not only depends on the currently detected respiratory data, but is also affected by the respiratory state at the previous moment. It can quickly capture sudden changes without causing misjudgment due to accidental noise fluctuations, providing robustness and flexibility for the calculation of respiratory activity indicators. By analyzing the difference between the respiratory activity index at the previous moment and the preset smoothing reference value, the size of the adaptive smoothing coefficient is adjusted, and a sensitivity coefficient is introduced to control the sensitivity of the adaptive smoothing coefficient to the respiratory activity index at the historical moment; Recursive memory can ensure that the historical state will not be obscured by short-term fluctuations. Specifically, it scales the respiratory activity index of the previous moment to ensure that it occupies a reasonable weight. At the same time, the scaling will weaken as the absolute value of the respiratory activity index at the historical moment increases, avoiding misleading extreme values. , in, Indicated in Respiratory activity indicators at all times; Represents the adaptive smoothing coefficient, which is used to dynamically adjust the current frame data and historical frame data (i.e. The weight of the respiratory activity index at the moment) in the calculation of the respiratory activity index to cope with individual differences in different breathing modes, the adaptive smoothing coefficient The specific calculation formula is: , in, represents the sensitivity coefficient, which is used to control the sensitivity of the adaptive smoothing coefficient to the respiratory activity index at the historical moment; Indicates the smoothing reference value, which is used to control the range of variation of the adaptive smoothing coefficient. It can be set according to the specific implementation scenario and is not limited here; Represents the base of the natural logarithm, which is used to construct the Sigmoid function so that the adaptive smoothing coefficient varies between 0 and 1; It means to sum the pixel changes at all pixel positions belonging to the breathing area matrix; Indicates at location The pixel change at reflects the degree of change of pixel values in two adjacent video frames; Represents the recursive memory coefficient, which is used to incorporate the respiratory activity index of historical moments into the calculation process of the respiratory activity index to avoid short-term noise interference. The specific calculation formula is: , in, Represents the weight coefficient, which is used to control the weight of the respiratory activity index at the historical moment; Indicated in The absolute value of the respiratory activity index at the time (i.e., the respiratory activity index at the historical moment); Represents the power exponent, which is used to control the nonlinearity of recursive memory; S3, based on the respiratory activity index, the respiratory frequency is calculated by a segmented hybrid transformation algorithm; Based on the respiratory activity index, the respiratory frequency is calculated using a segmented hybrid transformation algorithm; the segmented hybrid transformation algorithm is designed to process complex respiratory patterns under different states during the respiratory process, and dynamically adapts to changes in the patient's respiratory state by combining quadratic functions, sine functions, logarithmic functions, and cosine functions to achieve personalized respiratory frequency calculation; Compare the current respiratory activity index with the preset respiratory activity strength threshold. If the current respiratory activity index is lower than the respiratory activity strength threshold, the patient is judged to be in a stable breathing state; if the current respiratory activity index is equal to or higher than the respiratory activity strength threshold, the patient's breathing is judged to be in a complex state, that is, there are signs of acceleration or irregularity; When the patient is in a stable breathing state, a smooth and gradual way is used to describe the breathing pattern, combining the secondary change of time and the periodic fluctuation. The secondary change of time is used to describe the gradual growth of the breathing signal, which is suitable for scenes with slow and regular breathing, such as the patient's breathing when sleeping or resting. A periodic sine wave fluctuation is also superimposed to further capture the slight fluctuations in breathing. The amplitude and frequency of the sine wave are controlled by specific parameters to reflect the subtle fluctuations in the individual's breathing state. When the patient's breathing is in a complex state, it automatically switches to another method suitable for complex breathing states, which uses a combination of logarithmic function and cosine function to describe sudden changes and irregular fluctuations in respiratory frequency. The logarithmic function gradually slows down with the increase of time, indicating that the patient's respiratory frequency may tend to be stable, while the cosine function is used to capture short-term and drastic fluctuations in respiratory frequency, which can accurately reflect the changes in the patient's respiratory rhythm, especially in the case of sudden respiratory abnormalities; The formula for calculating respiratory rate is: , in, Indicated in The breathing rate at each moment is used to determine whether the breathing is normal; Indicated in Respiratory activity indicators at all times; Indicates time The derivative is used to calculate the rate of change of the respiratory activity index over time, thereby obtaining the respiratory rate; It represents the product of the piecewise transformation function and the respiratory activity index, accurately reflecting the current respiratory rate; Indicated in The piecewise transformation function of the respiratory activity at each moment is used to capture the changing trend of the respiratory pattern under different states and reflect the stable or fluctuating state of the patient's breathing. The specific calculation formula of the piecewise transformation function is: , in, and represent the first weight parameter and the second weight parameter, respectively, corresponding to different breathing states, Acting on the quadratic term, it is used to reflect the intensity of respiratory fluctuations when the patient is in a stable breathing state. Acting on logarithmic terms, it is used to reflect complex or irregular breathing conditions; and Respectively represent the first amplitude coefficient and the second amplitude coefficient, which are used to adjust the amplitude of the sine function and the cosine function respectively. Used to control the fluctuation amplitude of the patient's stable breathing. Used to control the vibration amplitude when the patient's breathing is unstable; and represent the first frequency parameter and the second frequency parameter, respectively, and are used to control the periods of the sine function and the cosine function, respectively; and Represent the sine function and cosine function, respectively, which are used to describe the periodic characteristics of respiratory fluctuations and simulate the natural periodic changes of breathing; Represents a logarithmic function, which is used to describe breathing changes under irregular conditions, adapt to gradual changes in breathing frequency, and is suitable for situations where breathing frequency changes gradually; Indicates the respiratory activity strength threshold, which is used to divide the respiratory activity state into two stages (stable respiratory state and complex state). It can be set according to the specific implementation scenario and is not limited here; By constantly comparing the current respiratory activity index with the respiratory activity strength threshold and switching between different states, it can flexibly respond to the real-time changes in the patient's breathing, ensuring the real-time and accuracy of respiratory rate calculation; S4. Calculate a personalized detection threshold through a personalized breathing abnormality detection algorithm; judge the breathing rate based on the personalized detection threshold to obtain a judgment result; and adopt an intervention strategy according to the judgment result; A personalized detection threshold is calculated by a personalized breathing abnormality detection algorithm, and the personalized breathing abnormality detection algorithm realizes real-time monitoring and abnormality detection of breathing patterns through a dual mechanism of dynamic threshold adjustment and abnormality detection judgment; Doctors set the basic detection threshold based on clinical experience as the default criterion for judging respiratory rate changes. In actual applications, respiratory patterns are affected by circadian rhythms. For example, respiratory rate decreases during sleep at night, while it increases during daytime activities. In order to match the patient's circadian rhythm, a threshold dynamic adjustment mechanism is adopted, and a physiological rhythm modulation item is introduced. The modulation coefficient of the physiological rhythm is set for different time periods, so that the personalized detection threshold fluctuates with the time period, and the fluctuation is periodic. The length of each cycle is consistent with the human body's circadian rhythm and is set to 24 hours to adapt to the natural regulation of breathing by the physiological cycle. Next, the current respiratory activity index will be compared with the historical average respiratory activity index. The comparison process relies on a recursive adaptive update mechanism. At the same time, a sensitivity parameter is introduced to dynamically adjust the detection accuracy according to the difference between the current respiratory activity index and the historical average respiratory activity index. When the current respiratory activity fluctuates violently compared with the historical state, the sensitivity parameter will automatically increase, and the personalized detection threshold for abnormalities will also decrease accordingly. When the respiratory state is stable, the sensitivity parameter will decrease to reduce the probability of false alarms. The historical average respiratory activity index comes from an existing database. The calculation formula for the personalized detection threshold is: , in, Indicated in Personalized detection threshold at each moment to determine whether the respiratory rate change is abnormal; Indicates the basic detection threshold, that is, the default respiratory rate change judgment standard without special adjustment, which is a fixed value set by doctors based on clinical experience; Represents the physiological rhythm modulation item, which is used to simulate the diurnal fluctuation of respiratory frequency and reflect the natural fluctuation of human respiratory frequency in different time periods. For example, breathing is stable at night and may increase during daytime activities. The modulation coefficient of the physiological rhythm is used to control the amplitude of the rhythm fluctuation; Represents a sine wave function, which is used to simulate the periodic changes of day and night. represents one complete cycle in radians, Indicates the length of the day and night cycle, set to 24 hours; Represents an adaptive adjustment item, which is used to dynamically adjust the personalized detection threshold; represents the exponential function part, which is used to quantify the difference between the current respiratory activity index and the historical average respiratory activity index as an adjustment factor of the personalized detection threshold; Indicates the historical average respiratory activity index; It represents the sensitivity parameter, which is used to control the influence of the difference between the current respiratory activity index and the historical average respiratory activity index on the threshold change; The abnormal detection judgment mechanism is used to compare the personalized detection threshold with the difference between the current respiratory rate and the historical average respiratory rate to obtain the judgment result, that is, if , then the output , indicating that an abnormality is detected, otherwise the output , indicating that the respiratory rate is within the normal range, the formula is as follows: , in, Indicated in The judgment result at each moment is used to determine whether there is abnormal breathing; Indicates the current breathing rate The historical average respiratory rate The absolute difference between the historical average respiratory rate and the From existing databases; The intervention strategies are as follows: , in, Indicates the medical intervention strategy. Intervene immediately when Observe regularly; By generating intervention strategies in a timely manner, the risk of respiratory abnormalities not being discovered in time is avoided and patient safety is improved.
[0025] The respiratory patient management system based on video surveillance realizes real-time monitoring of the patient's respiratory status, and allows long-term observation and data analysis without wearing any equipment. While reducing patient discomfort, it also reduces the cost of equipment use and the workload of medical staff, and improves the convenience and coverage of medical monitoring.
[0026] In summary, a respiratory patient management system and method based on video surveillance has been completed.
[0027] The order of the embodiments of the invention is for description only and does not represent the advantages and disadvantages of the embodiments. The processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.
[0028] The various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referenced to each other, and each embodiment focuses on the differences from other embodiments.
[0029] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that the technical solutions described in the above embodiments may still be modified, or some of the technical features may be replaced by equivalents. Such modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention, and should be included in the protection scope of the present invention.
Claims
1. A method for managing respiratory patients based on video monitoring, characterized in that: The following steps are involved: S1, real-time acquisition of patient images to obtain a video frame pixel matrix; Based on the video frame pixel matrix, the pixel change and the breathing area matrix are calculated by the breathing area detection and extraction algorithm; S2, based on the pixel change and the respiratory area matrix, the respiratory activity index is calculated by an adaptive recursive smoothing algorithm; S3, based on the respiratory activity index, the respiratory frequency is calculated by a segmented hybrid transformation algorithm; S4, calculating a personalized detection threshold through a personalized breathing abnormality detection algorithm; Based on the personalized detection threshold, the respiratory rate is judged to obtain the judgment result; Adopt intervention strategies based on the judgment results.
2. A respiratory patient management method based on video monitoring according to claim 1, characterized in that: The S1 specifically includes: In the process of implementing the breathing area detection and extraction algorithm, nonlinear compression and Gaussian weight enhancement are introduced to obtain the pixel change amount. The specific calculation formula of the pixel change amount is: , in, Indicates at location The amount of pixel change at ; represents the hyperbolic tangent function used to compress the magnitude of pixel changes; Indicates at location Gaussian weight at ; Indicates the height of the video frame; Indicates the width of the video frame; represents the standard deviation of the Gaussian function; express Always at the location The pixel value at ; express Always at the location The pixel value at ; A threshold screening mechanism is introduced to obtain the breathing region matrix.
3. A respiratory patient management method based on video monitoring according to claim 1, characterized in that: The S2 specifically includes: The adaptive recursive smoothing algorithm introduces an adaptive smoothing coefficient and a recursive memory coefficient to calculate the respiratory activity index; the specific calculation formula is: , in, Indicated in Respiratory activity indicators at all times; represents the adaptive smoothing coefficient; represents the breathing zone matrix; Indicates at location The amount of pixel change at ; Indicated in Respiratory activity indicators at all times; Represents the recursive memory coefficient.
4. A respiratory patient management method based on video monitoring according to claim 3, characterized in that: The S2 specifically includes: The sensitivity coefficient is introduced, and the adaptive smoothing coefficient is adjusted in combination with the respiratory activity index at the previous moment.
5. A respiratory patient management method based on video monitoring according to claim 3, characterized in that: The S2 specifically includes: The weight coefficient is introduced to scale the respiratory activity index at the previous moment to obtain the recursive memory coefficient.
6. A respiratory patient management method based on video monitoring according to claim 1, characterized in that: The S3 specifically includes: In the process of implementing the segmented hybrid transformation algorithm, the segmented transformation function is introduced to dynamically adapt to the changes in the patient's respiratory state, calculate the rate of change of the respiratory activity index over time, and obtain the respiratory frequency; the specific formula of the segmented transformation function is: , in, Indicated in Piecewise transformation function of respiratory activity at each moment; and denote the first weight parameter and the second weight parameter, respectively, corresponding to different breathing states; and represent a first amplitude coefficient and a second amplitude coefficient, respectively, and are used to adjust the amplitudes of the sine function and the cosine function, respectively; and represent the first frequency parameter and the second frequency parameter, respectively, and are used to control the periods of the sine function and the cosine function, respectively; Indicates the threshold of respiratory activity strength.
7. A respiratory patient management method based on video monitoring according to claim 1, characterized in that: The S4 specifically includes: In the process of implementing the personalized breathing abnormality detection algorithm, physiological rhythm modulation items and adaptive adjustment items are introduced to obtain personalized detection thresholds. Based on the personalized detection thresholds, the breathing pattern is monitored in real time and abnormality detection is performed. The calculation formula of the personalized detection threshold is: , in, Indicated in Personalized detection threshold at each moment; represents the basic detection threshold; represents the circadian rhythm modulation term; represents the modulation coefficient of the physiological rhythm; Indicates the length of the day-night cycle; Represents an adaptive adjustment item; Indicates the historical average respiratory activity index; Represents the sensitivity parameter.
8. A respiratory patient management system based on video surveillance, applied to a respiratory patient management method based on video surveillance as claimed in claim 1, characterized in that: Includes the following sections: Video acquisition module, respiratory area detection and extraction module, respiratory activity index calculation module, respiratory rate calculation module, personalized abnormality detection and intervention module; The video acquisition module acquires patient images in real time and obtains a video frame pixel matrix; Output the video frame pixel matrix to the breathing area detection and extraction module; The breathing area detection and extraction module calculates the pixel change amount and the breathing area matrix based on the video frame pixel matrix through the breathing area detection and extraction algorithm; and outputs the pixel change amount and the breathing area matrix to the breathing activity index calculation module; The respiratory activity index calculation module calculates the respiratory activity index based on the pixel change and the respiratory area matrix through an adaptive recursive smoothing algorithm; Outputting the respiratory activity index to a respiratory rate calculation module; The respiratory rate calculation module calculates the respiratory rate based on the respiratory activity index through a segmented hybrid transformation algorithm; Output the respiratory rate to the personalized abnormality detection and intervention module; The personalized anomaly detection and intervention module calculates the personalized detection threshold based on the respiratory rate through the personalized respiratory anomaly detection algorithm; Based on the personalized detection threshold, the current respiratory rate is judged to obtain the judgment result, and an intervention strategy is adopted according to the judgment result.
Citation Information
Patent Citations
Real-time medical video image denoising method
CN102014240A
Pig respirator rate detecting method and device on basis of machine vision
CN102499687A
Robust heart rate estimation
CN107708538A
Feature transformation method and device based on segmentation statistics
CN108921183A
Non-contact apnea detection device and non-contact apnea detection method
CN111544001A
Cited By
AI-based breathing abnormity identification method and system
CN119517440A
Intensive care unit image processing method and system based on respiratory movement
CN119889666A
A respiratory motion-based intensive care unit image processing method and system
CN119889666B
