A postoperative monitoring system and method for cardiac surgery based on image processing

Abstract

The present invention relates to the field of image processing, and in particular to a post-cardiac surgery monitoring system and method based on image processing. The system comprises: obtaining a chest image of a patient, preprocessing the chest image of the patient by an improved image preprocessing algorithm, and obtaining a preprocessed chest image; based on the preprocessed chest image, using a healing progress evaluation algorithm to calculate an incision healing progress evaluation value, and setting an incision healing speed threshold and a standard progress value of the incision healing progress, and determining whether there is an abnormality in the incision healing progress; based on the preprocessed chest image, using an inflammation evaluation algorithm to calculate an inflammation severity evaluation value, and setting an inflammation severity threshold, and determining whether there is an inflammation abnormality in the incision. The system solves the existing technical problems of being difficult to accurately process small and frequent displacements, resulting in loss of edge information; being unable to fully combine spatial and temporal information, and being prone to ignoring the development trend of inflammation.

Description

Technical Field

[0001] The present invention relates to the field of image processing, and in particular to an image processing-based postoperative monitoring system and method for cardiac surgery. Background Art

[0002] With the advancement of medical technology, the success rate of modern cardiac surgery has significantly improved, but postoperative complications and risks are still the focus that medical staff need to closely monitor. Traditional postoperative monitoring methods for cardiac surgery rely on the monitoring of various physiological indicators, such as electrocardiogram, blood pressure, blood oxygen saturation, respiratory rate, etc. Although they can provide real-time physiological information, they cannot directly reflect changes in external visual features and require high-frequency observation and recording. Especially for critically ill patients, medical staff need to frequently check vital signs and physical conditions, which not only consumes a lot of human resources, but is also prone to problems such as inadequate monitoring or negligence.

[0003] With the rapid development of image processing technology, artificial intelligence and computer vision, postoperative monitoring systems and methods based on image processing have gradually become an important technological trend. Image processing technology can capture the patient's physical representation information in real time through the camera, and identify potential risk factors through algorithm analysis, thereby providing more sophisticated and comprehensive support for postoperative monitoring of cardiac surgery.

[0004] However, existing post-cardiac surgery monitoring methods have the following technical problems: it is difficult to accurately process small and frequent displacements, such as image blur caused by breathing or slight shaking; removing noise can easily lead to blurred details and loss of edge information; relying on single time point information of the image, it is impossible to capture the dynamic changes in the healing process; the inflammatory and swollen areas are difficult to distinguish in the RGB color space, resulting in low detection accuracy; it is impossible to fully combine spatial and temporal information, and the development trend of inflammation is easy to ignore. Summary of the invention

[0005] The present invention provides a postoperative monitoring system and method for cardiac surgery based on image processing to solve the existing technical problems that it is difficult to accurately process small and frequent displacements, such as image blur caused by breathing or slight shaking; the removal of noise easily leads to blurred details and loss of edge information; the reliance on single time point information of the image cannot capture the dynamic changes in the healing process; the inflammatory and swollen areas are difficult to distinguish in the RGB color space, resulting in low detection accuracy; the spatial and temporal information cannot be fully combined, and the development trend of inflammation is easily ignored.

[0006] The present invention provides a post-cardiac surgery monitoring system and method based on image processing, which specifically includes the following technical solutions:

[0007] A postoperative monitoring method for cardiac surgery based on image processing comprises the following steps:

[0008] S1: Acquire a chest image of a patient, and preprocess the chest image of the patient using an improved image preprocessing algorithm to obtain a preprocessed chest image;

[0009] S2: Based on the preprocessed chest image, a healing progress evaluation algorithm is used to calculate an incision healing progress evaluation value, and an incision healing speed threshold and a standard progress value of the incision healing progress are set to determine whether there is an abnormality in the incision healing progress;

[0010] S3: When there is an abnormality in the incision healing progress, the inflammation severity assessment value is calculated based on the preprocessed chest image using the inflammation assessment algorithm, and the inflammation severity threshold is set to determine whether there is an inflammatory abnormality in the incision.

[0011] Preferably, the S1 specifically includes:

[0012] The improved image preprocessing algorithm includes a motion compensation part and a noise removal and edge preservation part. The motion compensation part calculates the displacement of each pixel in the chest image by analyzing the pixel position difference between adjacent frames, and corrects the position of the pixels in the chest image.

[0013] Preferably, the S1 specifically includes:

[0014] The noise removal and edge preservation part calculates the noise distribution according to the local noise characteristics of each pixel in the chest image, and takes filtering operations of different intensities for areas with different noise distributions. A nonlinear diffusion model is introduced to perform selective diffusion according to the characteristics of the pixel area. The pixel intensity of the preprocessed chest image is obtained by calculating the gradient of the pixel points in the chest image. The calculation formula is:

[0015] ,

[0016] in, Indicated in The spatial coordinates at that moment are The pixel intensity of the preprocessed chest image; Indicated in The spatial coordinates at that moment are The pixel intensity of the patient's chest image; and It represents the displacement compensation parameter caused by the slight movement of the patient during the imaging process. Compensation in Direction of movement, Compensation in Direction of movement; represents the first adjustment coefficient; Indicates that the chest image is at position and Noise distribution at time instant; represents the second adjustment coefficient; Represents the second-order change rate of the local chest image; Indicates that the chest image is Direction and The first derivative in the direction.

[0017] Preferably, the S2 specifically includes:

[0018] In the implementation of the healing progress evaluation algorithm, the incision edge is detected by calculating the transverse and longitudinal gradients of the preprocessed chest image, and the noise interference is suppressed by combining the local topological perturbation to obtain the edge strength of the preprocessed chest image. The specific calculation formula is:

[0019] ,

[0020] in, Indicated in The spatial coordinates at that moment are The edge strength of the preprocessed chest image; Indicates the pixel intensity of the chest image after preprocessing Directional gradient; Indicates the pixel intensity of the chest image after preprocessing Directional gradient; Represents the square root operation; and represents the square term; represents the exponential function; Represents the adjustment parameter that controls the influence of the local topological perturbation on the gradient calculation results; represents the amount of local topological perturbation.

[0021] Preferably, the S2 specifically includes:

[0022] In the implementation of the healing progress evaluation algorithm, by setting the incision area detection threshold, the edge intensity of the preprocessed chest image is compared with the set incision area detection threshold, the pixels that meet the conditions are extracted, and the complete incision contour is constructed.

[0023] Preferably, the S2 specifically includes:

[0024] In the implementation of the healing progress evaluation algorithm, after the incision area is detected, an adaptive weighting coefficient is introduced to adjust the weight value of the pixel point; the incision healing progress is evaluated by calculating the rate of change of the incision edge over time; and the incision healing progress evaluation value is obtained by calculating the second-order derivative of the incision edge with respect to time.

[0025] Preferably, the S3 specifically includes:

[0026] In the implementation process of the inflammation assessment algorithm, the preprocessed chest image is subjected to color conversion to obtain the red channel value of the pixel in the chest image after color separation; after color separation, the inflammation area is detected by the extracted red channel value, the inflammation area detection threshold is set, and the red channel value is compared with the set inflammation area detection threshold to extract the pixels that meet the conditions to form a pixel set of the inflammation area.

[0027] Preferably, the S3 specifically includes:

[0028] In the process of implementing the inflammation assessment algorithm, spatial derivative adjustment parameters and time derivative adjustment parameters are introduced to analyze the inflammation area in spatial and temporal dimensions to obtain the inflammation severity assessment value.

[0029] A postoperative monitoring system for cardiac surgery based on image processing, comprising the following parts:

[0030] Image acquisition module, image preprocessing module, healing progress assessment module, inflammation detection and assessment module, monitoring and early warning module;

[0031] Image acquisition module: acquires the patient's chest image and outputs the patient's chest image to the image preprocessing module;

[0032] Image preprocessing module: Based on the patient's chest image from the image acquisition module, the patient's chest image is preprocessed by an improved image preprocessing algorithm to obtain a preprocessed chest image, and the preprocessed chest image is output to the healing progress assessment module and the inflammation detection and assessment module;

[0033] Healing progress assessment module: Based on the chest image preprocessed by the image preprocessing module, the healing progress assessment value is calculated using the healing progress assessment algorithm, the healing speed threshold and the standard progress value of the incision healing progress are set, and whether there is abnormal incision healing is determined, and the abnormal healing determination result is output to the monitoring and early warning module;

[0034] Inflammation detection and assessment module: Based on the chest image preprocessed by the image preprocessing module and the healing abnormality determination result of the monitoring and early warning module, the inflammation severity assessment value is calculated using the inflammation assessment algorithm, the inflammation severity threshold is set, and it is determined whether there is inflammation abnormality in the incision, and the inflammation abnormality determination result is output to the monitoring and early warning module;

[0035] Monitoring and early warning module: receives the abnormal healing judgment results from the healing progress assessment module, notifies medical staff to handle and outputs the abnormal healing judgment results to the inflammation detection and assessment module, and receives the abnormal inflammation judgment results from the inflammation detection and assessment module.

[0036] The beneficial effects of the technical solution of the present invention are:

[0037] 1. By analyzing the pixel position differences between adjacent frames, the displacement of each pixel in the patient's chest image is calculated, and accurate compensation is performed in the horizontal and vertical directions to correct image distortion caused by involuntary movements such as breathing and slight tremors. This effectively eliminates the problem of chest image blurring caused by slight movements of the patient, keeps the chest image at high resolution and continuity, and avoids chest image distortion affecting subsequent judgments.

[0038] 2. According to the local noise distribution of pixels, dynamic filtering operations are used to remove noise. Through the nonlinear diffusion model, diffusion is strengthened in flat areas to eliminate noise, and diffusion is weakened in edge areas to retain details. The Laplace operator is combined to enhance edge information to avoid the problem of blurred details during the denoising process.

[0039] 3. By detecting the brightness changes of pixels, the edge pixels of the incision are identified, and a complete incision contour is constructed based on the local topological disturbance. An adaptive weighting coefficient is further introduced to evaluate the incision healing state according to the smoothness of pixel changes. The second-order derivative is used to capture the acceleration or deceleration trend of the incision healing progress. The incision healing speed threshold and the standard progress value of the incision healing progress are used to determine whether there is any abnormality in the incision healing progress. The incision healing progress is accurately evaluated, timely feedback is provided, manual intervention is reduced, and the efficiency of postoperative monitoring is improved.

[0040] 4. By converting the chest image from RGB color space to HSV color space, separating the red channel information, highlighting the characteristics of the inflammatory area, it is easy to identify the inflammatory area and reduce misjudgment. The detection threshold of the inflammatory area is set, and the pixels that meet the conditions are extracted to construct a pixel point set of the inflammatory area. By analyzing the spatial texture and time change trend of the inflammatory area, the signal of worsening or relief of the disease is captured in time, and the accuracy of the evaluation process is controlled by adjusting the spatial derivative parameters and the time derivative parameters. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] Figure 1 This is a structural diagram of a post-cardiac surgery monitoring system based on image processing according to the present invention;

[0042] Figure 2 This is a flow chart of a post-cardiac surgery monitoring method based on image processing described in the present invention. DETAILED DESCRIPTION

[0043] 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.

[0044] 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.

[0045] The specific scheme of a post-surgery monitoring system and method based on image processing provided by the present invention is described in detail below with reference to the accompanying drawings.

[0046] See attached Figure 1 , which shows a structural diagram of a postoperative cardiac surgery monitoring system based on image processing provided by an embodiment of the present invention, the system includes the following parts:

[0047] Image acquisition module, image preprocessing module, healing progress assessment module, inflammation detection and assessment module, monitoring and early warning module;

[0048] Image acquisition module: uses a camera to obtain the patient's chest image and outputs the patient's chest image to the image preprocessing module;

[0049] Image preprocessing module: Based on the patient's chest image from the image acquisition module, the patient's chest image is preprocessed by an improved image preprocessing algorithm to obtain a preprocessed chest image, and the preprocessed chest image is output to the healing progress assessment module and the inflammation detection and assessment module;

[0050] Healing progress assessment module: Based on the chest image preprocessed by the image preprocessing module, the healing progress assessment value is calculated using the healing progress assessment algorithm, the healing speed threshold and the standard progress value of the incision healing progress are set to determine whether there is an incision healing abnormality, and the healing abnormality determination result is output to the monitoring and early warning module;

[0051] Inflammation detection and assessment module: Based on the chest image preprocessed by the image preprocessing module and the healing abnormality determination result of the monitoring and early warning module, the inflammation severity assessment value is calculated using the inflammation assessment algorithm, and the inflammation severity threshold is set to determine whether there is inflammation abnormality in the incision, and the inflammation abnormality determination result is output to the monitoring and early warning module;

[0052] Monitoring and early warning module: receives the abnormal healing judgment results from the healing progress assessment module, notifies medical staff for processing and outputs the abnormal healing judgment results to the inflammation detection and assessment module for further evaluation of whether there is abnormal inflammation in the incision, and receives the abnormal inflammation judgment results from the inflammation detection and assessment module, notifies medical staff for processing.

[0053] See attached Figure 2 , which shows a flow chart of a postoperative monitoring method for cardiac surgery based on image processing provided by an embodiment of the present invention, the method comprising the following steps:

[0054] S1, obtaining a chest image of a patient, and preprocessing the chest image of the patient by using an improved image preprocessing algorithm to obtain a preprocessed chest image;

[0055] Using a camera to obtain a chest image of a patient, preprocessing the chest image of the patient using an improved image preprocessing algorithm to obtain a preprocessed chest image;

[0056] The improved image preprocessing algorithm includes a motion compensation part and a noise removal and edge preservation part, which are used to remove noise and compensate for image blur caused by slight movement of the patient, while maintaining image details and edge information;

[0057] The motion compensation part calculates the displacement of each pixel in the chest image by analyzing the pixel position difference between adjacent frames. The displacement is respectively the displacement in the horizontal and vertical directions. After determining the displacement, the position of each pixel in the chest image is corrected, that is, the chest image is corrected, and the chest image distortion caused by the patient's slight movement during the postoperative monitoring period is corrected, which can not only avoid the blurring problem of the chest image, but also ensure the continuity of the chest image information;

[0058] The noise removal and edge preservation part calculates the noise distribution of each pixel in the chest image according to the local noise characteristics of each pixel, and takes filtering operations of different intensities for areas with different noise distributions; at the same time, a nonlinear diffusion model is introduced to ensure that the noise can be removed while retaining the edge information, and the noise is removed by gradually smoothing the pixels in the chest image, but not all pixels are smoothed in the same way, but selective diffusion is performed according to the characteristics of the area where the pixels are located, and the gradient of each pixel point in the chest image is calculated, that is, the change of the pixel value with the spatial position, in the area with extremely small gradient, that is, the flat area of ​​the chest image (such as a large area of ​​the same color), the diffusion will be strengthened, thereby effectively removing the noise, and in the area with extremely large gradient, that is, the edge area or the area with rich details, the diffusion will be weakened to ensure that the details will not be blurred due to denoising; it is particularly suitable for medical image processing, because in postoperative monitoring, the detailed information of the image (such as the outline of the incision and the change of inflammation) is crucial and must be retained;

[0059] The calculation formula of the pixel intensity of the preprocessed chest image is:

[0060] ,

[0061] in, Indicated in The spatial coordinates at that moment are The pixel intensity of the preprocessed chest image; Indicated in The spatial coordinates at that moment are The pixel intensity of the patient's chest image; and It represents the displacement compensation parameter caused by the slight movement of the patient during the imaging process, which is used to reduce the image blur caused by the patient's involuntary movements (such as breathing, slight tremor), help image correction and restore a clear chest image. Compensation in Direction of movement, Compensation in Direction of movement; Represents the first adjustment coefficient, used for trade-off and the nonlinear diffusion term The impact on chest image preprocessing determines the balance between noise removal and detail preservation; Indicates that the chest image is at position and The noise distribution at the time instant is used to help remove the influence of random noise on the details of the chest image; Represents the second adjustment coefficient, which is used to control the influence of image gradient in processing, balance the weight of edge enhancement and detail retention, and prevent edge blurring caused by over-smoothing; The Laplace operator represents the local second-order rate of change of the chest image, which is used to enhance details and edges, retain significant image details and edge structures, and especially keep edge information from being lost during noise removal. Indicates that the chest image is Direction and The first-order derivative in the direction, namely the image gradient, describes the rate of change of pixel intensity in the chest image and is used to identify the contours and details in the image; Indicates the normalization of image gradients to control the gradient influence of the image;

[0062] S2. Based on the preprocessed chest image, a healing progress evaluation algorithm is used to calculate an incision healing progress evaluation value, and an incision healing speed threshold and a standard progress value of the incision healing progress are set to determine whether there is an abnormality in the incision healing progress;

[0063] Based on the preprocessed chest image, the healing progress assessment value of the incision is calculated using the healing progress assessment algorithm;

[0064] The healing progress evaluation algorithm detects the incision edge by calculating the transverse and longitudinal gradients of the preprocessed chest image, and suppresses noise interference by combining the local topological disturbance amount to ensure the accuracy of edge detection; a comprehensive evaluation of the incision healing progress is achieved by analyzing the incision edge changes in different time periods, and the incision healing progress evaluation value is calculated, which not only considers the edge information in the spatial direction, but also combines the time analysis to ensure the comprehensiveness and accuracy of the evaluation; the edge point of the incision is judged by checking the brightness changes of each pixel point along the horizontal and vertical directions, and if the brightness of the pixel is Direction and If there is a significant change in the direction, then the pixel point may be the edge point of the incision. A calculation method based on the local topological perturbation amount is further introduced. The local topological perturbation amount is used to measure the brightness difference between the pixel point and the surrounding neighboring pixels. By continuously comparing the difference values ​​between the pixel point and the surrounding pixels, the outline and boundary of the incision are gradually detected.

[0065] The calculation formula of the edge strength of the preprocessed chest image is:

[0066] ,

[0067] in, Indicated in The spatial coordinates at that moment are The edge intensity of the preprocessed chest image reflects the edge change and is used to detect the contour of the incision; Indicates the pixel intensity of the chest image after preprocessing Directional gradient, reflecting lateral edge information; Indicates the pixel intensity of the chest image after preprocessing Directional gradient, reflecting vertical edge information; represents the square root operation, which ensures the positive value of the gradient; and represents the square term, which is used to eliminate the positive and negative differences in the gradient direction, ensuring that the edge strength measurement of the preprocessed chest image only reflects the gradient size and is not affected by the direction; represents an exponential function, which weights the gradient value by adjusting the amount of local topological disturbance, suppresses insignificant edge changes, and avoids noise interference or erroneous detection of non-target areas; Represents the adjustment parameter used to control the influence of the local topological perturbation on the gradient calculation result; Represents the amount of local topological disturbance, used to measure the target pixel exist The brightness difference between the pixels at the moment and the neighboring pixels reflects the significant changes in the local area and is used to identify the edge of the cut. The calculation formula for the local topological disturbance is:

[0068] ,

[0069] in, Indicates the target pixel The surrounding neighborhood pixel set can be specifically set according to the specific implementation scenario and is not limited here; Represents a set of pixels in the neighborhood Pixel coordinates in ; Represents the neighborhood pixel set The brightness differences of all pixels in the image are accumulated to capture the overall changes in the target pixel area. Indicates the absolute value sign, eliminates positive and negative differences, and ensures that the calculated brightness difference reflects the changes at the edge of the cut; Indicated in The spatial coordinates at that moment are The pixel intensity of the preprocessed chest image;

[0070] By setting the incision area detection threshold, the edge strength of the preprocessed chest image is compared with the set incision area detection threshold, all pixels that meet the conditions are extracted, and then a complete incision contour is constructed. The formula is as follows:

[0071] ,

[0072] in, Indicated in The set of all pixels that meet the conditions at the moment is the detected incision area; Indicated in The threshold used to detect the incision area at any time can be set according to the specific implementation scenario and is not limited here; Indicates that all conditions are met A set of pixels;

[0073] After the incision area is detected, in order to further accurately reflect the recovery of the incision, an adaptive weighting coefficient is introduced to adjust the weight value of each pixel. If the brightness change of the pixel becomes smooth, it means that it is in a healing state, and the pixel is given a very low weight; on the contrary, if the brightness of the pixel changes significantly, it is considered that the healing is not yet completed, and the pixel is given a very high weight.

[0074] In order to capture the change in speed during the incision healing process, further analyze the change of the incision boundary over time, and evaluate whether the incision healing progress is accelerated or slowed down by calculating the change rate of the incision edge over time, it can not only track the overall progress of incision healing, but also determine whether the healing speed of the incision is normal within a certain period of time; specifically, it is achieved by calculating the second-order derivative of the incision edge with respect to time, and the second-order derivative reflects the acceleration or deceleration trend of incision healing;

[0075] The calculation formula for the wound healing progress assessment value is:

[0076] ,

[0077] in, Indicated in The incision healing progress assessment value at the moment; Indicates the summation operation of all pixels belonging to the cutout area; represents the adaptive weighting coefficient, which adjusts the influence of different incision areas on the healing progress; represents a balance parameter, which is used to adjust the local smoothness of the pixel intensity of the preprocessed chest image; The Laplacian operator representing the pixel intensity of the preprocessed chest image reflects the degree of local changes around the pixel point; represents the pixel intensity of the chest image after preprocessing; It represents the second-order derivative of the edge intensity of the preprocessed chest image with respect to time, which is used to represent the acceleration of the incision healing process and reflect the changing trend of the incision healing progress; Indicated in The spatial coordinates at that moment are The edge intensity of the preprocessed chest image reflects the edge change;

[0078] If the incision healing progress changes slowly and is lower than the standard progress value of the incision healing progress, it indicates that the incision healing is abnormal, otherwise, the incision healing is normal;

[0079] The specific implementation formula is as follows:

[0080] ,

[0081] in, The first-order derivative of the wound healing progress evaluation value with respect to time reflects the rate of change of the wound healing progress and is used to judge the speed of wound healing; Indicates the incision healing speed threshold, which can be set according to the specific implementation scenario and is not limited here; Standard progress value indicating the progress of wound healing;

[0082] When there is an abnormality in the healing progress of the incision, it is necessary to notify the medical staff to handle it and evaluate whether there is abnormal inflammation;

[0083] S3. When there is an abnormality in the healing progress of the incision, the inflammation severity assessment value is calculated based on the pre-processed chest image using the inflammation assessment algorithm, and the inflammation severity threshold is set to determine whether there is abnormal inflammation in the incision;

[0084] When there is an abnormality in the healing progress of the incision, it is necessary to further consider whether there is inflammation in the incision; based on the pre-processed chest images, the inflammation severity assessment value is calculated using the inflammation assessment algorithm;

[0085] The inflammation assessment algorithm can quickly locate the inflammation area and judge its severity through color space conversion separation, inflammation area detection and inflammation severity assessment based on spatial and temporal derivatives;

[0086] The input preprocessed chest image is converted from the RGB (red, green, blue) color space to the HSV (hue, saturation, brightness) color space. The RGB color space is a traditional color representation method. However, in the RGB color space, different combinations of colors make it difficult to clearly distinguish the red characteristics of the inflammation area. The HSV color space is suitable for analyzing the characteristics of colors, especially red, by separating the brightness and saturation of colors. Through a standard color conversion process, the data of each pixel point is mapped from the RGB color space to the HSV color space to highlight the information of the red channel.

[0087] The color conversion process formula is as follows:

[0088] ,

[0089] in, Represents the pixel points in the chest image after color separation exist The red channel value at the moment is obtained by converting the preprocessed chest image from the RGB color space to the HSV color space, which is used to detect the inflammation area;

[0090] After color separation, the inflammatory area is detected through the extracted red channel, and the inflammatory area detection threshold is set. exist The red channel value at the moment is compared with the set inflammation area detection threshold, and all pixels that meet the conditions are extracted to form a pixel set of the inflammation area. The formula is as follows:

[0091] ,

[0092] in, Indicated in The set of all pixels that meet the conditions at the moment is the detected inflammation area; Indicated in The threshold used to detect the inflammation area at any time can be set according to the specific implementation scenario and is not limited here; Indicates that all conditions are met A set of pixels;

[0093] To assess the severity of inflammation, all detected pixels are traversed to calculate texture changes and time changes, aiming to analyze the development of inflammation in two dimensions: one is the change in the spatial dimension (i.e., redness and swelling of adjacent areas or local exudation); the other is the change in the temporal dimension (i.e., whether inflammation is expanding or relieving over time); the change detection in the spatial dimension reveals the dramatic fluctuations in texture, such as sudden redness and swelling, by calculating the intensity of local changes between pixels. At the same time, in order to avoid misjudgment caused by local changes in details, a spatial derivative adjustment parameter is designed to control the weight of spatial changes in the evaluation process. The change detection in the temporal dimension is used to analyze the trend of each pixel in the time series to ensure that the rapid expansion or relief of inflammation can be captured;

[0094] The calculation formula for the assessment value of inflammation severity is:

[0095] ,

[0096] in, Indicated in The moment-by-moment inflammation severity assessment value is used to quantify the severity of the detected inflammation area; It represents the sum of all pixels belonging to the inflammation area; Represents the spatial derivative adjustment parameter, which is used to control the weight of the spatial second-order derivative, adjust the impact of spatial changes on the evaluation results, and avoid misjudgment caused by changes in local details; Represents pixel The spatial second-order derivative of is used to detect the intensity of local texture changes, reflecting texture changes such as redness or exudation; represents the time derivative adjustment parameter, which is used to control the weight of the time-varying derivative to ensure that the changes in the development speed of inflammation can be detected; Represents pixel The time derivative of reflects the rate of change of the inflammatory area over time;

[0097] If the inflammation severity assessment value exceeds the inflammation severity threshold, it indicates that there is an inflammation abnormality, otherwise, it is normal;

[0098] ,

[0099] in, Indicates the threshold of inflammation severity, which can be set according to the specific implementation scenario and is not limited here;

[0100] Accurate reminders of inflammatory abnormalities can easily prompt doctors to take further intervention.

[0101] In summary, a postoperative monitoring system and method for cardiac surgery based on image processing has been completed.

[0102] 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.

[0103] 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.

[0104] 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 postoperative monitoring method for cardiac surgery based on image processing, characterized in that: The following steps are involved: S1: Acquire a chest image of a patient, and preprocess the chest image of the patient using an improved image preprocessing algorithm to obtain a preprocessed chest image; The improved image preprocessing algorithm includes a motion compensation part and a noise removal and edge preservation part. The motion compensation part calculates the displacement of each pixel in the chest image by analyzing the pixel position difference between adjacent frames, and corrects the position of the pixels in the chest image. The noise removal and edge preservation part calculates the noise distribution according to the local noise characteristics of each pixel in the chest image, and takes filtering operations of different intensities for areas with different noise distributions; A nonlinear diffusion model is introduced to perform selective diffusion according to the characteristics of the pixel area. The pixel intensity of the preprocessed chest image is obtained by calculating the gradient of the pixel points in the chest image. The calculation formula is: ， in, Indicated in The spatial coordinates at that moment are The pixel intensity of the preprocessed chest image; Indicated in The spatial coordinates at that moment are The pixel intensity of the patient's chest image; and It represents the displacement compensation parameter caused by the slight movement of the patient during the imaging process. Compensation in Direction of movement, Compensation in Direction of movement; represents the first adjustment coefficient; Indicates that the chest image is at position and Noise distribution at time instant; represents the second adjustment coefficient; Represents the second-order change rate of the local chest image; Indicates that the chest image is Direction and The first derivative in direction; S2: Based on the preprocessed chest image, the healing progress evaluation algorithm is used to calculate the incision healing progress evaluation value. In the implementation process of the healing progress evaluation algorithm, the incision edge is detected by calculating the transverse and longitudinal gradients of the preprocessed chest image, and the local topological perturbation amount is combined to suppress noise interference to obtain the edge strength of the preprocessed chest image. The specific calculation formula is: ， in, Indicated in The spatial coordinates at that moment are The edge strength of the preprocessed chest image; Indicates the pixel intensity of the chest image after preprocessing Directional gradient; Indicates the pixel intensity of the chest image after preprocessing Directional gradient; Represents the square root operation; and represents the square term; represents the exponential function; Represents the adjustment parameter that controls the influence of the local topological perturbation on the gradient calculation results; represents the amount of local topological perturbation; By setting the incision area detection threshold, the edge strength of the preprocessed chest image is compared with the set incision area detection threshold, the pixels that meet the conditions are extracted, and a complete incision outline is constructed; After the incision area is detected, an adaptive weighting coefficient is introduced to adjust the weight value of the pixel point; the incision healing progress is evaluated by calculating the change rate of the incision edge over time; and the incision healing progress evaluation value is obtained by calculating the second-order derivative of the incision edge with respect to time; A threshold value of the incision healing speed and a standard progress value of the incision healing progress are set to determine whether there is an abnormality in the incision healing progress; S3: When there is an abnormality in the incision healing progress, the inflammation severity assessment value is calculated based on the preprocessed chest image using the inflammation assessment algorithm, and the inflammation severity threshold is set to determine whether there is an inflammatory abnormality in the incision.

2. The postoperative monitoring method for cardiac surgery based on image processing according to claim 1, characterized in that: The S3 specifically includes: In the implementation process of the inflammation assessment algorithm, the preprocessed chest image is subjected to color conversion to obtain the red channel value of the pixel in the chest image after color separation; after color separation, the inflammation area is detected by the extracted red channel value, the inflammation area detection threshold is set, and the red channel value is compared with the set inflammation area detection threshold to extract the pixel points that meet the conditions to form a pixel point set of the inflammation area.

3. The postoperative monitoring method of cardiac surgery based on image processing according to claim 2, characterized in that: The S3 specifically includes: In the process of implementing the inflammation assessment algorithm, the spatial derivative adjustment parameters and the time derivative adjustment parameters are introduced to analyze the inflammation area in the spatial and temporal dimensions to obtain the inflammation severity assessment value. The specific calculation formula is: ， in, Indicated in Assessment value of inflammation severity at each moment; It represents the sum of all pixels belonging to the inflammation area; represents the spatial derivative adjustment parameter; Represents pixel The spatial second-order derivative of ; represents the time derivative adjustment parameter; Represents pixel The time derivative of .

4. A post-cardiac surgery monitoring system based on image processing, applied to the post-cardiac surgery monitoring method based on image processing according to claim 1, characterized in that: Includes the following sections: Image acquisition module, image preprocessing module, healing progress assessment module, inflammation detection and assessment module, monitoring and early warning module; Image acquisition module: acquires the patient's chest image and outputs the patient's chest image to the image preprocessing module; Image preprocessing module: Based on the patient's chest image from the image acquisition module, the patient's chest image is preprocessed by an improved image preprocessing algorithm to obtain a preprocessed chest image, and the preprocessed chest image is output to the healing progress assessment module and the inflammation detection and assessment module; Healing progress assessment module: Based on the chest image preprocessed by the image preprocessing module, the healing progress assessment value is calculated using the healing progress assessment algorithm, the healing speed threshold and the standard progress value of the incision healing progress are set, and whether there is abnormal incision healing is determined, and the abnormal healing determination result is output to the monitoring and early warning module; Inflammation detection and assessment module: Based on the chest image preprocessed by the image preprocessing module and the healing abnormality determination result of the monitoring and early warning module, the inflammation severity assessment value is calculated using the inflammation assessment algorithm, the inflammation severity threshold is set, and it is determined whether there is inflammation abnormality in the incision, and the inflammation abnormality determination result is output to the monitoring and early warning module; Monitoring and early warning module: receives the abnormal healing judgment results from the healing progress assessment module, notifies medical staff to handle and outputs the abnormal healing judgment results to the inflammation detection and assessment module, and receives the abnormal inflammation judgment results from the inflammation detection and assessment module.

Images