Abdominal wall local anesthetic diffusion visual analysis system based on infrared temperature image features

Abstract

The invention discloses an abdominal wall local anesthetic diffusion visual analysis system based on infrared temperature image features, and relates to the technical field of medical monitoring. The thermal imaging acquisition and processing module is used for performing sequence scanning at preset time points set before and after retardation operation to obtain original temperature image data; the target region feature extraction module is used for determining a target analysis region according to the retardation type and determining feature parameters according to the temperature data of the target analysis region; and the effect prediction and evaluation module takes the characteristic parameter set as the input of a time sequence prediction model, and outputs and obtains the corresponding predicted low-temperature area and predicted retardation effect level. Through the infrared thermal imaging technology, the temperature change caused by diffusion of the local anesthetic in the abdominal wall tissue after injection can be continuously captured in a non-contact and non-radiation mode, so that the two-dimensional diffusion profile and the dynamic migration process of the local anesthetic can be visually presented in real time, and then corresponding visual information can be provided.

Description

Technical Field

[0001] This invention relates to the field of medical monitoring technology, specifically to a visualization analysis system for the diffusion of local anesthetics in the abdominal wall based on infrared temperature image features. Background Technology

[0002] Regional nerve blocks of the abdominal wall (such as transversus abdominis plane block, rectus abdominis sheath block, etc.) are an important regional anesthesia technique that has been widely used for perioperative analgesia in abdominal surgery. The core of this technique lies in the precise injection of local anesthetics into the target fascial space, effectively reducing postoperative pain and decreasing the dosage of opioids and their related adverse reactions by blocking the nerve conduction innervating the abdominal wall.

[0003] Currently, in clinical practice, the assessment of the diffusion range and effectiveness of local anesthetics within the abdominal wall fascia mainly relies on the following methods:

[0004] Ultrasound-guided injection: The operator inserts the puncture needle into the target space under real-time ultrasound guidance and judges the diffusion by observing the "water separation" phenomenon of the drug solution during injection (i.e., the expansion of the low-echo area).

[0005] Clinical assessment: The diffusion of the drug solution can be indirectly inferred by observing the extent of sensory blockade (such as the area of ​​reduced pain sensation during acupuncture), the degree of motor blockade, or the analgesic effect during or after the procedure.

[0006] Electrical nerve stimulation: mainly used to locate nerves, with limited help in assessing drug diffusion.

[0007] Computed tomography (CT) or magnetic resonance imaging (MRI): Although they can provide high-resolution images of anatomy and drug distribution, they are limited to research scenarios and are not suitable for routine clinical procedures due to their high cost, complex operation, radiation exposure (CT), long processing time, and inability to be used in real time.

[0008] Chinese invention patent CN120510152A discloses an AI-based intelligent management system for information on multiple chronic diseases. This system, equipped with a laser response capture module and a multi-dimensional data processing module, can collect RGB image datasets and infrared thermogram datasets of the skin area at set time points after laser treatment. Based on an AI image processing model, the images are standardized and the vitiligo region is segmented. Morphological parameters such as the area of ​​the vitiligo patches, edge curvature kurtosis, and tangential perturbation, as well as thermal parameters such as temperature gradient recovery value, temperature change curvature, and regional heat flux density, are then extracted. The system utilizes a structural convergence analysis module and a thermal recovery analysis module working together to calculate the structural convergence tension index (STCI) and the thermal response consistency mapping index (HRCI). This comprehensively measures the skin response process from both structural and thermodynamic dimensions, effectively overcoming the inaccuracies caused by traditional methods relying on visual observation and subjective judgment.

[0009] However, the above and similar technical solutions still have the following shortcomings: In clinical practice of performing nerve block techniques such as rectus abdominis sheath block, although ultrasound guidance is the gold standard, it can only show the needle tip position and the local anechoic area at the initial stage of drug injection. It cannot visualize the two-dimensional diffusion contour and three-dimensional distribution volume of local anesthetic in the tissue in real time and comprehensively, and it is also difficult to monitor its dynamic diffusion process. Furthermore, the method of assessing the block effect by relying on the patient's subjective feedback or changes in vital signs during the operation is not only highly subjective and easily interfered with, but also cannot be applied to special patient groups such as sedated, comatose, or patients with communication difficulties. Summary of the Invention

[0010] The purpose of this invention is to provide a visualization analysis system for the diffusion of local anesthetics in the abdominal wall based on infrared temperature image features, so as to solve the problems mentioned in the background art.

[0011] To achieve the above objectives, the present invention provides the following technical solution: a visualization analysis system for the diffusion of local anesthetics in the abdominal wall based on infrared temperature image features, comprising:

[0012] Thermal imaging acquisition and processing module: Performs sequential scanning at preset time points before and after the blocking operation to obtain raw temperature image data;

[0013] Target area feature extraction module: Based on the type of blockage, it determines the target analysis region and, based on the temperature data corresponding to the target analysis region, determines the feature parameters, including:

[0014] SA1: Target area definition: The type of blockage is determined through the graphical user interface, and each pixel in the original temperature image data is identified and classified according to the type of blockage to determine the target analysis area.

[0015] SA2: Temperature difference calculation: The temperature value corresponding to each pixel in the infrared thermal image before the blockage is used as the reference temperature value. The temperature value corresponding to each pixel in the target analysis area is compared with the reference temperature value to determine the temperature difference of each pixel in the target analysis area. According to the set color mapping rules, the temperature difference distribution map corresponding to the target analysis area is obtained.

[0016] SA3: Feature Extraction: Based on the temperature difference distribution map, a set of feature parameters is determined, including static parameters and dynamic parameters;

[0017] Effect prediction and evaluation module: The feature parameter set is used as the input of the time series prediction model, and the corresponding predicted low temperature zone area and predicted blocking effect level are output.

[0018] Furthermore, the preset time points include, but are limited to, before block, before puncture after disinfection, 5 minutes after block, 10 minutes after block, 15 minutes after block, 20 minutes after block, and 30 minutes after block.

[0019] Furthermore, the original temperature image data is preprocessed using an image preprocessing module to obtain a preprocessed infrared thermal image, including:

[0020] W1: Temperature Correction: The ambient temperature and relative humidity data corresponding to each original temperature image data are collected by an ambient temperature and humidity sensor. The corresponding correction coefficient is determined by a set calibration parameter lookup table. At the same time, the radiation value corresponding to each pixel in the original temperature image data is corrected according to the correction coefficient to determine the corresponding corrected radiation value and obtain the temperature-corrected infrared thermal image.

[0021] W2: Motion Compensation: The infrared thermal image before lag is used as the reference image, and the remaining infrared thermal image is used as the image to be registered. The feature points in the reference image and the image to be registered are matched by a feature point comparison algorithm to select the corresponding correct matching points. At the same time, a mathematical transformation model is constructed based on the correct matching points to determine the transformation parameters. Based on the transformation parameters, the coordinates of each pixel in the image to be registered are transformed to obtain the preprocessed infrared thermal image.

[0022] Furthermore, the ambient temperature and humidity sensor, infrared thermal imager, and blackbody source are all placed in a climate chamber. By adjusting the ambient temperature in the climate chamber in real time, data pairs including ambient temperature value, original radiation value, and actual blackbody source temperature value are collected. At the same time, based on the data pairs, the corresponding conversion formula is fitted and obtained. Based on the slope and intercept of the conversion formula, the corresponding correction coefficient is determined.

[0023] Furthermore, a mathematical transformation model is obtained through the affine transformation model and the correct matching points, and the corresponding transformation parameters are determined based on the mathematical transformation model, including translation components, scaling and rotation components, and shearing and rotation components.

[0024] Furthermore, anatomical landmarks on the body surface are displayed by superimposing semi-transparent or dashed lines, and the target analysis area is determined by auxiliary delineation within the anatomical landmarks. The anatomical landmarks include, but are not limited to, the virtual projection lines of the linea alba, the lower edge of the costal arch, the anterior superior iliac spine, and the midaxillary line.

[0025] Furthermore, based on the temperature difference of each pixel in the temperature difference distribution map, static parameters are determined, including:

[0026] SB1: Low-temperature region area: Based on the comparison between the temperature difference corresponding to each pixel in the temperature difference distribution map and the preset temperature difference threshold, the number of pixels in the low-temperature region is determined, and the area of ​​the low-temperature region is determined based on the actual physical area of ​​a single pixel.

[0027] SB2: Average temperature drop amplitude: Based on the temperature difference corresponding to each pixel in the low-temperature region, the total temperature difference in the low-temperature region is determined, and based on the number of pixels in the low-temperature region, the average temperature drop amplitude is determined;

[0028] SB3: Spatial distribution morphology: The centroid and length of the low-temperature region are used as inputs to the constructed morphology determination model, and the corresponding spatial distribution morphology of the low-temperature region is obtained as output.

[0029] Furthermore, the coordinates of each pixel in the low-temperature region are weighted and averaged to obtain the average pixel coordinates of the low-temperature region, thereby determining the centroid coordinates of the low-temperature region.

[0030] Furthermore, a two-dimensional data point set is constructed based on the coordinates of each pixel in the low-temperature region, and the first principal component direction and the second principal component direction are determined by the principal component analysis algorithm. The first principal component direction and the second principal component direction are perpendicular to each other. At the same time, the first principal component direction is the major axis direction angle of the low-temperature region, and the second principal component direction is the minor axis direction angle of the low-temperature region.

[0031] Furthermore, based on the temperature difference between each pixel in the low-temperature region, dynamic parameters are determined, including:

[0032] SC1: Area expansion rate: Based on the infrared thermograms taken before puncture, 5 minutes after blockade, 10 minutes after blockade, 15 minutes after blockade, 20 minutes after blockade, and 30 minutes after blockade, the area of ​​the low-temperature region corresponding to each time point is determined. Based on the time point corresponding to each area of ​​the low-temperature region, an area-time series data pair is constructed. At the same time, the corresponding fitting line is obtained through linear fitting, and the corresponding average area expansion rate is determined.

[0033] SC2: Cooling front movement speed: The edge detection algorithm analyzes and detects the temperature difference distribution map of each frame to determine the boundary of the low temperature region in each frame. At the same time, the low temperature region boundaries of adjacent frames are compared to determine the corresponding displacement vector. Based on the time interval between adjacent frames of temperature difference distribution maps, the average movement speed of the cooling front is determined.

[0034] SC3: Cooling Core Migration Trajectory: Based on the temperature difference of each pixel in the temperature difference distribution map of each frame, the minimum temperature difference in the temperature difference distribution map of each frame is determined, and the pixels corresponding to the minimum temperature difference are connected in chronological order according to the time point corresponding to the temperature difference distribution map of each frame to obtain the cooling core migration trajectory.

[0035] Compared with the prior art, the beneficial effects of the present invention are:

[0036] Firstly, this invention uses infrared thermal imaging technology to continuously capture the temperature changes caused by the diffusion of local anesthetics in the abdominal wall tissue after injection, i.e. the low-temperature region, in a non-contact and radiation-free manner. This allows for the real-time and intuitive presentation of the two-dimensional diffusion profile and dynamic migration process of the local anesthetic, thereby providing corresponding visualization information.

[0037] Secondly, by extracting relevant static and dynamic parameters, this invention transforms the assessment of the blocking effect, which originally relied on the doctor's subjective experience and the patient's subjective feedback, into an objective and quantitative analysis based on data, thereby reducing errors and uncertainties caused by subjective judgment.

[0038] Thirdly, the present invention, through the constructed time-series prediction model, can predict the future low-temperature zone area and the final level of blockade effect based on the temperature diffusion characteristics collected in the early stage. This can help determine the success rate of blockade, and when the prediction effect is not good, remedial measures can be taken in time, thereby improving the success rate of analgesia and optimizing the patient's prognosis. Attached Figure Description

[0039] Figure 1 This is a system block diagram of the abdominal wall local anesthetic diffusion visualization analysis system of the present invention;

[0040] Figure 2 This is the infrared thermogram before the blocking in this invention;

[0041] Figure 3 This is an infrared thermogram taken 15 minutes after the blockage in this invention;

[0042] Figure 4 This is a temperature difference distribution diagram in this invention. Detailed Implementation

[0043] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0044] In clinical practice, while ultrasound guidance is the gold standard for nerve block techniques such as rectus abdominis sheath block, it can only display the needle tip position and the local anechoic area at the initial stage of drug injection. It cannot visualize the two-dimensional diffusion profile and three-dimensional distribution volume of the local anesthetic in the tissue in real time and comprehensively, nor can it monitor its dynamic diffusion process. Methods relying on patient subjective feedback or intraoperative vital sign changes to assess the block effect are not only highly subjective and susceptible to interference, but also unsuitable for sedated, comatose, or communication-impaired patients. The technical solution in this application uses an infrared thermal imager to perform sequential scanning at preset time points before and after the block procedure to acquire raw temperature image data of the abdominal wall region. Temperature correction and motion compensation are then applied to the raw temperature image data. Simultaneously, a target analysis area is defined according to the block type, and the temperature difference between the target analysis area and the area before the block is obtained, generating a corresponding temperature difference distribution map from which static and dynamic parameters are extracted. Simultaneously, static and dynamic parameters are used as inputs to the time-series prediction model, and the corresponding predicted low-temperature zone area and predicted blocking effect level are obtained as outputs, thereby realizing objective and visualized dynamic monitoring and effect evaluation of the local anesthetic diffusion process.

[0045] Example 1

[0046] refer to Figures 1-4 This embodiment provides a visualization analysis system for the diffusion of local anesthetics in the abdominal wall based on infrared temperature image features. This system includes a thermal imaging acquisition and processing module, a target area feature extraction module, and an effect prediction and evaluation module. The thermal imaging acquisition and processing module uses an infrared thermal imager to perform sequential scanning at preset time points before and after the blocking operation, acquiring the corresponding raw temperature image data. The target area feature extraction module extracts the corresponding static and dynamic parameters based on the temperature data of the target analysis area. The effect prediction and evaluation module uses the acquired static and dynamic parameters as input to a constructed time-series prediction model, outputting the predicted low-temperature zone area and the predicted blocking effect level.

[0047] In this embodiment, the thermal imaging acquisition and processing module is used to perform non-contact scanning of multiple preset time phases before, during, and after the abdominal wall nerve block procedure using an infrared thermal imager, and to acquire corresponding sequential infrared thermal images. It is worth noting that during the non-contact scanning of multiple time phases, at least seven infrared thermal images corresponding to different time points are acquired, including but not limited to before the block, before puncture after disinfection, 5 minutes after the block, 10 minutes after the block, 15 minutes after the block, 20 minutes after the block, and 30 minutes after the block.

[0048] Furthermore, in this embodiment, the infrared thermal imager is fixed to a floor stand or rail stand via a gimbal, and the height and angle of the fixed floor stand or rail stand can be adjusted, thus ensuring the infrared thermal imager remains stable and vibration-free during image data acquisition. It is worth noting that during stand adjustment, the center of the infrared thermal imager's lens must be perpendicular to the center of the patient's anterior abdominal wall, and the shooting distance must be 1-1.5 meters.

[0049] During the specific implementation process, when acquiring infrared thermal images before the block, the patient's anterior abdominal wall skin needs to be exposed for at least 15 minutes to allow the skin temperature to stabilize and reach room temperature before acquiring the corresponding pre-block infrared thermal image. Simultaneously, after disinfecting the patient's abdominal wall skin, the corresponding post-disinfection pre-puncture infrared thermal image is acquired. Furthermore, during ultrasound-guided puncture and local anesthetic injection, the injection time is timed in real time, and infrared thermal images are acquired at 5 minutes, 10 minutes, 15 minutes, 20 minutes, and 30 minutes after the block.

[0050] In this embodiment, the target area feature extraction module is used to determine the corresponding target analysis region based on the type of hindrance, and simultaneously extract the corresponding static and dynamic parameters based on the temperature data corresponding to the target analysis region. Specifically:

[0051] Step SA1: Target Region Definition. This involves defining the corresponding blockade type through a graphical user interface, including but not limited to rectus abdominis sheath block, transversus abdominis plane block, and iliohypogastric / ilioinguinal nerve block. Based on the defined blockade type, a deep learning semantic segmentation model, such as the U-Net architecture, is used to automatically match the corresponding analysis target region. In other words, the deep learning semantic segmentation model identifies and classifies each pixel in the infrared thermal image acquired by the thermal imaging acquisition and processing module to determine the corresponding target analysis region.

[0052] Furthermore, in this embodiment, the corresponding target analysis area can also be determined through auxiliary delineation. Specifically, in the infrared thermal image acquired by the thermal imaging acquisition and processing module, key anatomical landmarks on the body surface are displayed in a semi-transparent or dashed form, including but not limited to the virtual projection lines of the linea alba, the lower edge of the costal arch, the anterior superior iliac spine, and the mid-axillary line. Simultaneously, based on the displayed anatomical landmarks, the corresponding target analysis area is drawn on the image. It is worth noting that when assisted in delineating the target analysis area, the shape and size of the target analysis area can be specifically set according to actual needs.

[0053] Step SA2: Temperature Difference Calculation. This involves using the pre-blocking infrared thermal image as a reference image; specifically, the temperature value corresponding to each pixel in the pre-blocking infrared thermal image is used as the reference temperature value. Simultaneously, based on the temperature values ​​corresponding to each pixel in the infrared thermal image of the target analysis area, the temperature values ​​of each pixel in the target analysis area are compared with the corresponding reference temperature values ​​to determine the temperature difference between each pixel in the target analysis area. Specifically:

[0054]

[0055] in: To analyze the temperature difference at time t corresponding to pixel (x, y) in the infrared thermal image of the target area, we need to determine the target region. To analyze the temperature value of pixel (x, y) in the infrared thermal image of the target area at time t. This represents the temperature value at pixel (x, y) in the infrared thermal image before the blockage.

[0056] Furthermore, based on the temperature difference corresponding to each pixel in the target analysis area, it is matched with the set color mapping rules to determine the color distribution corresponding to each temperature difference, thereby obtaining the temperature difference distribution map corresponding to the target analysis area.

[0057] In the specific implementation process, the color mapping rules set in this embodiment are as follows: when the temperature difference is less than -2℃, the corresponding color distribution is dark blue; when the temperature difference is between -2℃ and -0.5℃, the corresponding color distribution is light blue or cyan; when the temperature difference is between -0.5℃ and +0.5℃, the corresponding color distribution is green; when the temperature difference is between +0.5℃ and +2℃, the corresponding color distribution is yellow or orange; and when the temperature difference is greater than +2℃, the corresponding color distribution is red.

[0058] Step SA3: Feature Extraction. Based on the temperature difference distribution map obtained in Step SA2, the corresponding set of feature parameters is determined, including static and dynamic parameters. Specifically, in this embodiment, the static parameters include the area of ​​the low-temperature region, the average temperature drop, and the spatial distribution pattern; the dynamic parameters include the area expansion rate, the speed of the cooling front movement, and the migration trajectory of the cooling core.

[0059] Furthermore, the static parameters in this embodiment include the area of ​​the low-temperature region, the average temperature drop amplitude, and the spatial distribution pattern. The area of ​​the low-temperature region is determined by comparing the temperature difference with a preset temperature difference threshold (which can be specifically set according to actual data requirements, and is therefore not specifically described in this embodiment). The average temperature drop amplitude is used to determine the average temperature difference of all pixels in the low-temperature region based on the temperature difference corresponding to each pixel. The spatial distribution pattern is used to specifically determine the shape of the low-temperature region based on its centroid and length.

[0060] Furthermore, the dynamic parameters in this embodiment include the area expansion rate, the cooling front movement speed, and the cooling core migration trajectory. The area expansion rate is used to linearly fit the low-temperature region area in consecutive time phases of the infrared thermograms taken before puncture, 5 minutes, 10 minutes, 15 minutes, 20 minutes, and 30 minutes after puncture following disinfection, to determine the corresponding instantaneous expansion rate. The cooling front movement speed is used to determine the low-temperature region boundary of each frame of the temperature difference distribution map using image processing algorithms (such as edge detection algorithms), and to determine the corresponding movement speed based on the average displacement and time interval of boundary pixels between adjacent frames. The cooling core migration trajectory is obtained by sequentially connecting the pixel coordinates corresponding to the lowest temperature in each frame of the temperature difference distribution map according to the corresponding time magnitude, thereby obtaining the corresponding trajectory line.

[0061] In this embodiment, the corresponding static parameters are determined based on the temperature difference corresponding to each pixel in the temperature difference distribution map, as follows:

[0062] Step SB1: Area of ​​the low-temperature region. This involves comparing the temperature difference corresponding to each pixel in the temperature difference distribution map with a preset temperature difference threshold (which can be set according to actual data requirements, and is not specifically described in this embodiment). Based on the comparison results, the corresponding low-temperature region is determined, specifically as follows:

[0063] When the temperature difference is greater than a preset temperature difference threshold, the area to which the corresponding pixel belongs is not a low-temperature area. Conversely, when the temperature difference is not greater than the preset temperature difference threshold, the area to which the corresponding pixel belongs is a low-temperature area.

[0064] Furthermore, based on the number of pixels corresponding to the identified low-temperature region and the actual physical area of ​​a single pixel, the number of pixels and the actual physical area of ​​a single pixel are combined to determine the area of ​​the corresponding low-temperature region.

[0065] Step SB2: Average temperature drop. Based on the low-temperature region identified in Step SB1, determine the temperature difference for each pixel within that region, and then determine the total temperature difference for the entire low-temperature region based on the temperature difference for each pixel.

[0066] Furthermore, by combining the total temperature difference corresponding to the low-temperature region with the number of pixels corresponding to the low-temperature region, the corresponding average temperature drop can be determined.

[0067] Step SB3: Spatial Distribution Pattern. This involves calculating a weighted average of the coordinates of each pixel within the low-temperature region to determine the average pixel coordinates, which are then the centroid coordinates of that region.

[0068] Furthermore, based on the coordinates of each pixel in the low-temperature region, a corresponding two-dimensional data point set is constructed. Simultaneously, principal component analysis (PCA) is used to analyze this data point set to determine the directions of the first and second principal components. The first principal component direction represents the most dispersed distribution of the two-dimensional data point set, which is the major axis direction of the low-temperature region. The second principal component direction is perpendicular to the first principal component direction, representing the minor axis direction of the low-temperature region.

[0069] Furthermore, based on the determined directions of the first and second principal components, the angle between the first principal component vector corresponding to the first principal component direction and the horizontal axis is determined, which is the corresponding major axis direction angle. Simultaneously, based on the second principal component vector corresponding to the second principal component direction, the ratio between the first and second principal component vectors is determined, thus obtaining the corresponding major and minor axis ratio.

[0070] In this embodiment, the obtained major axis direction angle and the target analysis target area determined in step SA1 are both used as inputs to the constructed morphological determination model (such as a rule-based expert system model or a fuzzy inference model), and the output is to obtain the corresponding spatial distribution morphology of the low-temperature region.

[0071] In this embodiment, the corresponding dynamic parameters are determined based on the temperature difference between each pixel in the low-temperature region, as follows:

[0072] Step SC1: Area expansion rate. Based on the infrared thermograms taken before puncture, 5 minutes after puncture, 10 minutes after puncture, 15 minutes after puncture, 20 minutes after puncture, and 30 minutes after puncture, the area of ​​the low-temperature region at each time point is determined. At the same time, based on each time point and the corresponding low-temperature region area, the corresponding area-time series data pairs are constructed.

[0073] Furthermore, based on the multiple area-time series data pairs obtained, a corresponding fitted line is obtained through linear fitting, and the corresponding average area expansion rate is determined based on the slope of the obtained fitted line.

[0074] Step SC2: Cooling Front Movement Speed. This involves analyzing and detecting the temperature difference distribution map in each frame using an edge detection algorithm, such as the Canny algorithm, to determine the boundaries of the low-temperature regions within each frame. Simultaneously, feature point matching or nearest neighbor algorithms are used to compare the determined low-temperature region boundaries in adjacent frames' temperature difference distribution maps, thereby determining the displacement vectors corresponding to the temperature difference distribution maps of adjacent frames.

[0075] Furthermore, based on the determined displacement vector, the corresponding average displacement magnitude is determined. Simultaneously, based on the time interval between adjacent temperature difference distribution maps, the determined average displacement and the time interval are combined to obtain the corresponding average moving speed, which is the average moving speed of the cooling front.

[0076] Step SC3: Cooling Core Migration Trajectory. This involves determining the minimum temperature difference in each frame's temperature difference distribution map based on the temperature difference between each pixel, and then determining the cooling core point in each frame's temperature difference distribution map based on the pixel corresponding to the minimum temperature difference.

[0077] Furthermore, based on the cooling core points and corresponding time points in each frame's temperature difference distribution map, the cooling core points are connected sequentially in chronological order to form the corresponding trajectory lines, i.e., the corresponding cooling core migration trajectories.

[0078] In this embodiment, the effect prediction and evaluation module is used to take the static and dynamic parameters obtained from the target area feature extraction module as inputs to the time series prediction model (such as the XGBoost model or RNN model), and output the corresponding predicted low temperature area and predicted blocking effect level.

[0079] Example 2

[0080] This embodiment provides a visualization analysis system for the diffusion of local anesthetics in the abdominal wall based on infrared temperature image features. The specific implementation method is the same as in Embodiment 1, except that the infrared thermal image acquired by the thermal imaging acquisition and processing module undergoes denoising, registration, and standardization processing to eliminate non-target interference in the infrared thermal image, resulting in a preprocessed infrared thermal image. The invention will be illustrated below with specific examples of this embodiment.

[0081] In this embodiment, the infrared thermal image obtained by the thermal imaging acquisition and processing module is preprocessed using an image preprocessing module to obtain a preprocessed infrared thermal image, as detailed below:

[0082] Step W1: Temperature Correction. During the acquisition of the corresponding infrared thermal image by the infrared thermal imager, the ambient temperature and relative humidity data at the corresponding time are collected using the accompanying ambient temperature and humidity sensors. Simultaneously, based on the collected ambient temperature and relative humidity data, these data are combined with a pre-defined calibration parameter lookup table to determine the corresponding correction coefficients, i.e., the corresponding gain and offset.

[0083] Furthermore, by determining the correction coefficients, namely gain and offset, the radiance value corresponding to each pixel in the original infrared thermal image is corrected to obtain the corresponding corrected radiance value. Based on the corrected radiance value, the corresponding temperature-corrected infrared thermal image is then obtained. In this embodiment, the radiance value corresponding to each pixel is corrected using gain and offset to obtain the corrected radiance value, specifically as follows:

[0084]

[0085] in: This is the corrected true temperature value. For gain, The original radiation value of the pixel. This is the offset.

[0086] Furthermore, an infrared thermal imager and a matching ambient temperature and humidity sensor are placed in a climate chamber containing a blackbody source. The temperature and humidity within the climate chamber are adjusted to collect the corresponding real temperature values. Based on the real temperature values ​​and the surface temperature of the blackbody source, a corresponding calibration parameter lookup table is established. In this embodiment, a blackbody source with a known surface emissivity close to 1 is placed inside the climate chamber. An infrared thermal imager and a matching ambient temperature and humidity sensor are positioned directly in front of the blackbody source. The ambient temperature within the climate chamber is adjusted in real time. At each stable ambient temperature, the ambient temperature and humidity sensor and the infrared thermal imager collect the corresponding ambient temperature and raw radiation values ​​within the climate chamber. In other words, by acquiring multiple data pairs (including ambient temperature, raw radiation, and the real temperature of the blackbody source), a corresponding conversion formula is fitted to obtain the corresponding conversion formula. Based on the obtained conversion formula, the corresponding correction coefficients are determined. Specifically, the slope in the conversion formula is the gain in the correction coefficients, and the intercept in the conversion formula is the offset in the correction coefficients.

[0087] Step W2: Motion Compensation. The infrared thermogram before the arrest is used as the reference image. Infrared thermograms taken before disinfection, 5 minutes after arrest, 10 minutes after arrest, 15 minutes after arrest, 20 minutes after arrest, and 30 minutes after arrest are used as the images to be registered. A feature point comparison algorithm (e.g., Euclidean distance detection) is used to match feature points in the reference image with feature points in the images to be registered, selecting the correct matching points. Simultaneously, based on the selected correct matching points, a corresponding mathematical transformation model is constructed to determine the corresponding transformation parameters, including translation, rotation angle, and scaling factor. In other words, based on the determined transformation parameters (i.e., translation components, scaling / rotation components, and shearing / rotation components), the coordinates of each pixel in the images to be registered are transformed. Based on the transformed pixel coordinates, a preprocessed infrared thermogram is obtained.

[0088] Furthermore, feature point detection algorithms (such as SIFT, ORB, and AKAZE) are used to identify each feature point in the reference image and the image to be registered, thus determining each feature point in both images. Simultaneously, feature point comparison algorithms are used to compare and match each feature point in the reference image and the image to be registered, thereby identifying multiple corresponding matching feature point pairs.

[0089] Furthermore, by using the affine transformation model and the obtained pairs of corresponding matching feature points, the corresponding transformation parameters, namely translation components, scaling and rotation components, and shearing and rotation components, are determined. In other words, based on the obtained translation, scaling and rotation, and shearing and rotation components, the coordinates of each pixel in the image to be registered are transformed, specifically as follows:

[0090]

[0091] in: These are the transformed column coordinates. These are the transformed row coordinates. These are the original column coordinates of the pixels. These are the original row coordinates of the pixels. , To scale the rotation component, , For the shear rotation component, , This represents the translation component.

[0092] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended embodiments and their equivalents.

Claims

1. An abdominal wall local anesthetic drug diffusion visualization analysis system based on infrared temperature image features, characterized by, The application relates to a method for predicting the effect of a blockage, comprising the following steps: a thermal imaging acquisition and processing module: a sequence scan is performed at a preset time point before and after the blockage operation, and original temperature image data is obtained; a target area feature extraction module: according to the blockage type, a target analysis area is determined, and a feature parameter is determined according to the temperature data corresponding to the target analysis area, comprising: SA1: target area definition: the blockage type is determined through a graphical user interface, and each pixel point in the original temperature image data is identified and classified according to the blockage type, so that a target analysis target area is determined; SA2: temperature difference calculation: the temperature value of each pixel point in the infrared thermal image before the blockage is taken as a reference temperature value, the temperature value of each pixel point in the target analysis target area is compared with the reference temperature value, the temperature difference of each pixel point in the target analysis target area is determined, and a temperature difference distribution map corresponding to the target analysis target area is obtained according to a set color mapping rule; SA3: feature extraction: according to the temperature difference distribution map, a feature parameter set is determined, comprising a static parameter and a dynamic parameter; an effect prediction and evaluation module: the feature parameter set is taken as the input of a time series prediction model, and a corresponding predicted low-temperature area and a predicted blockage effect grade are output.

2. The system for visualizing diffusion of local abdominal wall anesthetics based on features of infrared thermal images according to claim 1, characterized in that, The preset time points include but are not limited to before the blockage, after the sterilization and before the puncture, 5 minutes after the blockage, 10 minutes after the blockage, 15 minutes after the blockage, 20 minutes after the blockage and 30 minutes after the blockage.

3. The system for visualizing diffusion of local abdominal wall anesthetics based on features of infrared thermal images according to claim 1, characterized in that, The original temperature image data is preprocessed through an image preprocessing module, and a preprocessed infrared thermal image is obtained, comprising: W1: temperature correction: through an environment temperature and humidity sensor, environment temperature data and relative humidity data corresponding to each original temperature image data are acquired, a correction coefficient is determined through a set calibration parameter lookup table, and the radiation value corresponding to each pixel point in the original temperature image data is corrected according to the correction coefficient, so that a corrected radiation value is determined, and a temperature-corrected infrared thermal image is obtained; W2: motion compensation: the infrared thermal image before the blockage is taken as a reference image, the remaining infrared thermal images are taken as to-be-registered images, feature points in the reference image and the to-be-registered images are matched through a feature point comparison algorithm, correct matching points are selected, a mathematical transformation model is constructed according to the correct matching points, transformation parameters are determined, and the coordinates of each pixel point in the to-be-registered images are transformed according to the transformation parameters, so that a preprocessed infrared thermal image is obtained.

4. The system for visualizing diffusion of local abdominal wall anesthetics based on features of infrared thermal images according to claim 3, characterized in that, The environment temperature and humidity sensor, the infrared thermal imager and the blackbody source are all arranged in a climate box, data pairs including an environment temperature value, an original radiation value and a blackbody source real temperature value are acquired through real-time adjustment of the environment temperature in the climate box, a conversion formula is fitted and acquired according to the data pairs, and the slope and intercept of the conversion formula are used to determine the correction coefficient.

5. The system for visualizing diffusion of local abdominal wall anesthetics based on features of infrared thermal images according to claim 3, characterized in that, A mathematical transformation model is obtained through an affine transformation model and the correct matching points, and corresponding transformation parameters are determined according to the mathematical transformation model, including translation components, scaling and rotation components and shearing and rotation components.

6. The system for visualizing diffusion of local abdominal wall anesthetics based on features of infrared thermal images according to claim 1, characterized in that, The target analysis target area is determined in the body surface anatomical markers, including but not limited to the virtual projection lines of the abdominal white line, the lower edge of the rib arch and the anterior superior iliac spine and the midaxillary line, by superimposed display of the body surface anatomical markers in the form of semi-transparency or dotted lines.

7. The system for visualizing diffusion of local abdominal wall anesthetics based on features of infrared thermal images according to claim 1, characterized in that, According to the temperature difference of each pixel point in the temperature difference distribution map, static parameters are determined, including: SB1: low-temperature area area: according to the comparison result between the temperature difference corresponding to each pixel point in the temperature difference distribution map and the preset temperature difference threshold, the pixel point number of the low-temperature area is determined, and according to the actual physical area of a single pixel point, the area of the low-temperature area is determined; SB2: average temperature drop amplitude: according to the temperature difference corresponding to each pixel point in the low-temperature area, the total temperature difference of the low-temperature area is determined, and according to the pixel point number of the low-temperature area, the average temperature drop amplitude is determined; SB3: spatial distribution form: the centroid and long and short sleeves of the low-temperature area are taken as the input of the constructed form determination model, and the corresponding low-temperature area spatial distribution form is output.

8. The system for visualizing diffusion of local abdominal wall anesthetics based on features of infrared thermal images according to claim 7, characterized in that, The coordinates of each pixel point in the low-temperature area are weighted and averaged to obtain the average pixel point coordinates of the low-temperature area, and the centroid coordinates of the low-temperature area are determined.

9. The system for visualizing diffusion of local abdominal wall anesthetics based on features of infrared thermal images according to claim 7, characterized in that, According to the coordinates of each pixel point in the low-temperature area, a two-dimensional data point set is constructed, and through principal component analysis algorithm, the first principal component direction and the second principal component direction are determined, and the first principal component direction and the second principal component direction are perpendicular to each other, and the first principal component direction is the long axis direction angle of the low-temperature area, and the second principal component direction is the short axis direction angle of the low-temperature area.

10. The system for visualizing diffusion of local abdominal wall anesthetics based on features of infrared thermal images according to claim 1, characterized in that, According to the temperature difference of each pixel point in the low-temperature area, dynamic parameters are determined, including: SC1: area expansion rate: according to the infrared thermographs before disinfection and after puncture, 5 minutes after blocking, 10 minutes after blocking, 15 minutes after blocking, 20 minutes after blocking and 30 minutes after blocking, the low-temperature area areas corresponding to each time point are determined, and the area-time sequence data pairs are constructed according to the time points corresponding to each low-temperature area area, and the fitting straight line is obtained through linear fitting, and the average area expansion rate is determined; SC2: cooling front moving speed: each frame of temperature difference distribution map is analyzed and detected through an edge detection algorithm to determine the boundary of the low-temperature area in each frame of temperature difference distribution map, and the boundaries of the low-temperature areas of adjacent frames are compared to determine the corresponding displacement vectors, and the average moving speed of the cooling front is determined according to the time interval between adjacent frames of temperature difference distribution map. SC3: Cooling core migration trajectory: according to the temperature difference of each pixel point in each frame temperature difference distribution diagram, the minimum temperature difference in each frame temperature difference distribution diagram is determined, and according to the time point corresponding to each frame temperature difference distribution diagram, the pixel point corresponding to the minimum temperature difference is connected in time sequence, and the cooling core migration trajectory is obtained.

Images