Abnormal region identification method and system for lumbar infrared thermal image

Abstract

The application belongs to the technical field of image recognition, and specifically provides a lumbar infrared thermal image thermal distribution abnormal region identification method and system. The method mainly identifies the lumbar thermal image anatomical landmarks accurately, positions the bilateral facet joint projection points of each lumbar segment, defines a micro thermal unit with the projection point as the center, extracts the hotspot aggregation degree, local thermal gradient value and segment asymmetry index of the segment, and finally judges whether the three features meet the preset conditions at the same time to identify the abnormal region more likely to be caused by deep facet joint dysfunction. The application effectively solves the feature confusion and misjudgment risk caused by the low signal-to-noise ratio of mixed thermal signals on the body surface and the overlapping of deep and shallow heat source characteristics in the prior art, realizes high specificity identification of the abnormal region more likely to be caused by deep facet joint dysfunction, and significantly improves the accuracy of the analysis conclusion.

Description

Technical Field

[0001] This application belongs to the field of image recognition technology, and specifically relates to a method and system for identifying abnormal thermal distribution areas in waist infrared thermal images. Background Technology

[0002] When using a mid-infrared thermal imager to examine the lumbar region, a common practice is to acquire an infrared thermal image of the subject's posterior lumbar region using the mid-infrared thermal imager. The initial screening for potential inflammatory areas or functional abnormalities is achieved primarily by identifying localized high-temperature areas in the image that are visible to the naked eye and exceed an empirical threshold, or by simply comparing whether the difference in the average temperature of symmetrical areas on both sides exceeds a certain threshold.

[0003] However, the infrared thermal signal on the human lumbar region is a comprehensive manifestation of heat generated by heat sources of different depths and natures under the skin after conduction through tissues. The diffuse heat caused by superficial soft tissues, which covers a large area, and the relatively localized inflammatory heat caused by dysfunction of the deep lumbar facet joints, often overlap in the projection area on the body surface. This results in the focal thermal characteristics reflecting the state of the facet joints being easily masked by the broader and stronger background thermal noise of the soft tissues, leading to a decrease in the signal-to-noise ratio. At the same time, normal bilateral body temperature fluctuations and environmental interference also objectively exist. Therefore, it is difficult to reliably distinguish whether the observed thermal abnormalities originate from deep facet joint dysfunction that needs attention, or from superficial soft tissues or other non-target factors, thus directly affecting the accuracy of the analytical conclusions. Summary of the Invention

[0004] This application provides a method and system for identifying abnormal thermal distribution areas in lumbar infrared thermograms. This effectively solves the risks of feature confusion and misjudgment caused by the low signal-to-noise ratio of mixed thermal signals on the body surface and the overlap of deep and shallow heat source features in the prior art. It achieves highly specific identification of abnormal areas that are more likely to originate from deep small joint dysfunction, and significantly improves the accuracy of the analysis conclusions.

[0005] To achieve the above objectives, this application adopts the following technical solution:

[0006] In a first aspect, this application provides a method for identifying abnormal thermal distribution areas in waist infrared thermal images, including:

[0007] Acquire mid- and far-infrared thermal images of the posterior lumbar region of the subject and identify key anatomical landmarks on the mid- and far-infrared thermal images.

[0008] Based on key anatomical landmarks, the projection points of the bilateral facet joints of each lumbar vertebral segment were identified in the mid- and far-infrared thermal images.

[0009] A micro-thermal unit is defined with the center of the projection point. The statistical characteristics of the temperature distribution of the micro-thermal unit are extracted, a standard thermal matrix indexed by lumbar vertebral segments is generated, and the hot spot concentration degree of the micro-thermal unit is calculated.

[0010] Symmetry analysis was performed on the temperature distribution statistical characteristics of the standard thermodynamic matrix to generate a segmental asymmetry index that reflects the thermal differences between the two small joints.

[0011] Based on the standard thermodynamic matrix and the segmental asymmetry index, the thermodynamic gradient of each small joint micro-thermal unit relative to its local background is calculated, generating local thermodynamic gradient values.

[0012] For each small joint's projection point, the system determines whether the preset conditions are met based on the segment asymmetry index, local thermal gradient value, and hotspot aggregation degree. If so, the micro-thermal unit defined by the small joint's projection point is identified as an abnormal region.

[0013] Secondly, this application provides a system for identifying abnormal thermal distribution areas in waist infrared thermal images, including:

[0014] Landmark Recognition Module: Used to acquire mid- and far-infrared thermal images of the posterior lumbar region of the subject and identify key anatomical landmarks on the mid- and far-infrared thermal images.

[0015] Coordinate mapping module: used to identify the projection points of the bilateral facet joints of each lumbar vertebral segment in mid- and far-infrared thermal images based on key anatomical landmarks on the body surface.

[0016] Feature extraction module: used to define micro-thermal units with the center of the projection point, extract the statistical features of temperature distribution of micro-thermal units, generate a standard thermal matrix indexed by lumbar segments, and calculate the hot spot concentration of micro-thermal units.

[0017] Symmetry Analysis Module: Used to perform symmetry analysis on the temperature distribution statistical characteristics of the standard thermodynamic matrix, and generate a segmental asymmetry index that reflects the thermal differences between the two small joints.

[0018] Gradient Calculation Module: Used to calculate the thermal gradient of each small joint's micro-thermal unit relative to its local background based on the standard thermal matrix and segmental asymmetry index, generating local thermal gradient values.

[0019] Anomaly Detection Module: For each small joint's projection point, based on the segment asymmetry index, local thermal gradient value, and hotspot concentration, it determines whether preset conditions are met. If so, the micro-thermal unit defined by the small joint's projection point is identified as an abnormal region.

[0020] Thirdly, this application provides a readable storage medium, comprising: computer program instructions stored in the readable storage medium, wherein the computer program instructions are read and executed by a processor to perform the steps of a method for identifying abnormal thermal distribution areas in a waist infrared thermal image.

[0021] The beneficial effects of this application are:

[0022] This application effectively solves the risks of feature confusion and misjudgment caused by the low signal-to-noise ratio of mixed thermal signals on the body surface and the overlap of deep and shallow heat source features in the prior art by adopting a scheme based on anatomical positioning to extract hotspot concentration, local thermal gradient and segmental asymmetry index, and to perform multi-feature fusion judgment. It achieves high specificity identification of abnormal areas that are more likely to originate from deep small joint dysfunction, and significantly improves the accuracy of the analysis conclusions.

[0023] Other features and advantages of this application will be set forth in the following description, and will be apparent in part from the description, or may be learned by practicing the application. The objectives and other advantages of this application may be realized and obtained by means of the structures pointed out in the description and the accompanying drawings. Attached Figure Description

[0024] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0025] Figure 1 A flowchart illustrating the method for identifying abnormal thermal distribution areas in waist infrared thermograms according to this application is shown.

[0026] Figure 2 A schematic diagram of the process for obtaining the segmental asymmetry index in this application is shown;

[0027] Figure 3 A schematic diagram of the process for generating local thermal gradient values ​​in this application is shown;

[0028] Figure 4 A schematic diagram of the process for obtaining the fusion coefficient in this application is shown. Detailed Implementation

[0029] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0030] When diffuse thermal activity in superficial soft tissues overlaps with focal thermal signals in deep small joints, it is easily affected by background noise, which can lead to misjudgment. To solve this problem, the inventors found that feature extraction and fusion are needed from multiple levels, including thermal morphology, spatial symmetry, and local contrast. During the research, it was realized that hotspot aggregation can effectively reflect the concentration of heat distribution, segmental asymmetry index can assess segmental differences under physiological fluctuations, and local thermal gradient value can improve the local signal-to-noise ratio of the target point. Therefore, a joint judgment based on these three features was proposed.

[0031] In some embodiments, such as Figure 1 As shown, this application provides a method for identifying abnormal thermal distribution areas in waist infrared thermal images, including:

[0032] S1. Obtain a mid- to far-infrared thermal image of the posterior lumbar region of the subject, and identify key anatomical landmarks on the mid- to far-infrared thermal image.

[0033] Use a mid-to-far infrared thermal imager to observe from behind the subject. An infrared image of the back of the waist was collected from a distance of meters, taken from a frontal view. This image recorded the temperature distribution information of various points on the body surface, i.e., a mid- and far-infrared thermogram.

[0034] On the acquired mid- and far-infrared thermal images, several key and stable bony anatomical landmarks were identified using image processing algorithms, including the highest points of the bilateral iliac crests and identifiable spinous processes.

[0035] S2. Based on key anatomical landmarks, identify the projection points of the bilateral facet joints of each lumbar vertebral segment in the mid- and far-infrared thermal images.

[0036] Using key anatomical landmarks as spatial references, and based on the geometric relationships of the lumbar vertebrae in standard human anatomy atlases, the surface projection points of the left and right facet joints of each lumbar vertebra segment from L1 to S1 are determined through calculation and mapping in the pixel coordinate system of the mid- and far-infrared thermogram.

[0037] S3. Define a micro-thermal unit at the center of the projection point, extract the statistical characteristics of the temperature distribution of the micro-thermal unit, generate a standard thermal matrix indexed by lumbar vertebral segments, and calculate the hot spot concentration degree of the micro-thermal unit.

[0038] Centered on the coordinates of each projection point, a fixed-size area is extracted from the mid- and far-infrared thermal image. This area is defined as a micro-thermal unit, and each micro-thermal unit corresponds to a specific anatomical point of a specific lumbar vertebral segment.

[0039] For all pixels within each micro-thermal unit, extract their temperature values, calculate the temperature distribution statistical features of each unit, and arrange the temperature distribution statistical features of all micro-thermal units according to their respective lumbar vertebral segments to form a structured table, namely the standard thermal matrix.

[0040] For each micro-thermal unit, an index reflecting the degree of concentration of high-temperature pixels is calculated based on the temperature distribution of its internal pixels, namely the hot spot concentration. The hot spot concentration can be used to reflect the difference in the distribution pattern between relatively localized heat generation and diffuse heat generation over a larger area.

[0041] S4. Perform symmetry analysis on the temperature distribution statistical characteristics of the standard thermodynamic matrix to generate a segmental asymmetry index that reflects the thermal differences between the two small joints.

[0042] For each lumbar vertebral segment in the standard thermal matrix, the temperature distribution statistical characteristics of the corresponding left and right micro-thermal units are extracted. Symmetry analysis is performed on these two values ​​to calculate a numerical value that quantifies the degree of thermal asymmetry between the left and right sides of the segment, namely the segmental asymmetry index. The segmental asymmetry index can be used to assess the thermal differences at the segmental level under the conditions of normal bilateral body temperature physiological fluctuations and environmental disturbances.

[0043] S5. Based on the standard thermodynamic matrix and segmental asymmetry index, calculate the thermodynamic gradient of each small joint's micro-thermal unit relative to its local background, and generate local thermodynamic gradient values.

[0044] For each small joint's microthermal unit, a region is defined around it as a local background. The temperature distribution statistical characteristics of the microthermal unit are compared with the temperature characteristics of the local background. Combined with the segmental asymmetry index of the lumbar vertebral segment to which the unit belongs, the local thermal gradient value of the microthermal unit of the small joint is obtained through fusion calculation. The local thermal gradient value is used to enhance the contrast of the target signal with its surrounding background thermal noise through local difference, reflecting the prominence of the small joint relative to its surrounding environment.

[0045] S6. For each small joint projection point, determine whether the preset conditions are met based on the segmental asymmetry index, local thermal gradient value, and hot spot concentration. If so, identify the micro-thermal unit defined by the small joint projection point as an abnormal area more likely to originate from deep small joint dysfunction.

[0046] For example, if a patient's three characteristic values ​​at the right facet joint of L4 all meet the preset conditions, it is therefore identified as an abnormal region.

[0047] In some embodiments, based on key anatomical landmarks, the spinous processes of each lumbar vertebral segment and the projection points of the bilateral facet joints are identified in mid- and far-infrared thermal imaging, including:

[0048] S21. Identify the interspinous space between the fourth and fifth lumbar vertebrae.

[0049] Among the key anatomical landmarks identified in S1, locate the highest points of the bilateral iliac crests. Connect these two points to obtain a virtual line. According to anatomical knowledge, the line connecting the highest points of the bilateral iliac crests usually passes through... The interspinous space is identified by using the coordinates of the midpoint of the line connecting the two lumbar vertebrae on the image. This is determined to be the interspinous space between the fourth and fifth lumbar vertebrae.

[0050] S22. Using the interspinous space as a reference, the projection points of the bilateral facet joints of the remaining lumbar vertebral segments are calculated according to the preset vertebral height intervals and lateral distances.

[0051] Set a vertebral height interval parameter This represents the estimated vertical distance between adjacent spinous processes of the lumbar vertebrae on the image; a lateral distance parameter is set. , representing the horizontal lateral distance between the projection point of the facet joint and the spinous process.

[0052] by The ordinate of the interspinous space point Based on the reference point, the ordinates of the reference points on the head side are as follows: , , , respectively corresponding , Equal segments; the ordinates of the caudal sides are as follows: , , , respectively corresponding , For equal segments, the x-coordinates of all reference points and the x-coordinates of the interspinous space points are... Consistent results were obtained, ultimately yielding the reference point coordinates for each lumbar vertebral segment. .

[0053] For the calculated reference point coordinates of each lumbar vertebral segment The coordinates of the surface projection point of its left small joint are: The coordinates of the surface projection point of the right facet joint are: .

[0054] In some embodiments, a micro-thermal unit is defined at the center of the projection point, and the statistical characteristics of the temperature distribution of the micro-thermal unit are extracted to generate a standard thermal matrix indexed by lumbar vertebral segments, including:

[0055] Sa31. Using each projection point as the center, a rectangular area of ​​fixed pixel size is extracted from the mid- and far-infrared thermal image as a micro-thermal unit.

[0056] For each projection point, its coordinates are... Centered on a pixel matrix in the mid- and far-infrared thermal image, a pixel with a side length of [value missing] is extracted. A square region of a pixel, defined as a micro-thermal unit of the projection point. , Representing lumbar vertebral segments, This represents the left or right side. The micro-thermal unit contains information about the local temperature field centered on that dissected point.

[0057] Sa32. Extract the temperature values ​​of all pixels within each micro-thermal unit and calculate the median temperature of that micro-thermal unit. The median temperature is less sensitive to noise points than the average temperature and better represents the central trend of the region.

[0058] Sa33. Arrange the median temperatures of all micro-thermal units according to their respective lumbar vertebral segments to generate a standard thermal matrix. The row indices of the matrix represent different lumbar vertebral segments, and the columns correspond to the left and right micro-thermal units within the same segment, forming a standard thermal matrix. It reflects the core characteristics of standardized and decentralized thermal distribution of the projection points of the bilateral facet joints of each segment in the entire lumbar spine region of the subject.

[0059] In some embodiments, calculating the hotspot concentration of a microthermal unit includes:

[0060] Sb31. For each microthermal unit, after calculating its median temperature, iterate through all L×L pixels within the unit and count the number of pixels whose internal temperature value is higher than the median temperature. , This reflects the number of hotspot pixels with temperatures higher than the baseline temperature of the region.

[0061] Sb32. Calculate the ratio of the number of pixels to the total number of pixels within the micro-thermal unit to obtain the hotspot clustering degree. , The range of values ​​is If the high-temperature pixels are evenly distributed or few in number, The value is relatively low; if high-temperature pixels are highly concentrated in a small area within the cell, forming a hotspot, The value will be close to Or higher, hotspot concentration is used to distinguish between diffuse temperature rise and localized high temperature points.

[0062] In some embodiments, a symmetry analysis is performed on the temperature distribution statistical characteristics of the standard thermodynamic matrix to generate a segmental asymmetry index reflecting the thermal differences between the two facet joints, including:

[0063] S41. For each lumbar vertebral segment, extract the median temperature of the micro-thermal units of the facet joints on both sides of the lumbar vertebral segment from the standard thermal matrix to obtain the median temperature on the left and right sides.

[0064] S42. Calculate the absolute difference between the median temperature on the left and the median temperature on the right. absolute difference The temperature difference between the two facet joint regions of this lumbar vertebral segment was quantified.

[0065] S43. Based on the median temperature on the left side, the median temperature on the right side, and the absolute difference, the significance of the difference is assessed to obtain the segmental asymmetry index, which is used to reflect whether the difference is physiological or pathological significant.

[0066] In some embodiments, such as Figure 2 As shown, based on the median temperature on the left side, the median temperature on the right side, and the absolute difference, a significance assessment of the difference was performed to obtain the segmental asymmetry index, including:

[0067] S431. Divide the larger of the left and right median temperatures by the smaller value to obtain the thermal background unevenness of the lumbar segment. , , This indicates that the temperature on both sides is completely balanced. The larger the value, the more significantly the temperature on one side is higher than the other, indicating an uneven background thermal environment.

[0068] The standardized relative difference R for the lumbar vertebral segment is obtained by calculating the ratio of the absolute difference to the sum of the median temperatures on the left and right sides. This is a dimensionless measure of relative difference, which reduces the influence of absolute temperature level on the perceived difference.

[0069] S432. Calculate the modulation factor based on the thermal background inhomogeneity and the standardized relative difference. , , The relative difference corresponding to the background imbalance degree of the representative unit, when the background imbalance degree When the temperature is very high, the temperature on one side is much higher than that on the other side, even if a certain relative difference is observed. The value will also be relatively small, suggesting that the difference may mainly stem from the imbalance in the background itself.

[0070] S433. Calculate the segmental asymmetry index based on standardized relative differences and modulation factors. , This represents the segmental asymmetry index.

[0071] denominator This constitutes an adjustment item, when A very small value indicates a balanced background or relatively reasonable differences. The segmental asymmetry index approaches the standardized relative difference when When the value is very large, the background is very uneven and the differences are significant. The value will be in the denominator Significant suppression. Therefore, relative differences that occur under a thermally balanced background tend to be assessed as more significant asymmetries, while the significance assessment of additional differences that occur under conditions of severe background imbalance is suppressed, thus helping to distinguish between global thermal shifts and true local asymmetries.

[0072] In some embodiments, such as Figure 3 As shown, based on the standard thermodynamic matrix and segmental asymmetry index, the thermodynamic gradient of each small joint's micro-thermal unit relative to its local background is calculated, generating local thermodynamic gradient values, including:

[0073] S51. For a projection point of a small joint, its micro-thermal unit is a square with a side length of L. In the mid- and far-infrared thermogram, based on the outer boundary of this square, it is expanded outward by w pixels to form a larger square ring. In this ring-shaped region, pixels covered by the micro-thermal units of other projection points are excluded. The area formed by all the remaining pixels is the local background region B of the projection point. The local background region represents the temperature of the surrounding tissue that is adjacent to the target point but is not directly affected by the thermal radiation of other identifiable anatomical points.

[0074] S52. Calculate the average temperature of all pixels in the local background area to obtain the average background temperature value.

[0075] S53. The median temperature of the micro-thermal unit, the background average temperature value, and the segment asymmetry index are fused to obtain the fusion coefficient, which is used to modulate the final gradient value.

[0076] S54. Based on the fusion coefficient, median temperature, and background average temperature, calculate the local thermal gradient value of the small joint. , , , and These represent the local thermal gradient value, fusion coefficient, median temperature, and background average temperature value of the micro-thermal unit, respectively.

[0077] Fusion coefficient As a multiplicative factor, when When the temperature difference is amplified, Time remains unchanged, when This will reduce the temperature difference.

[0078] In some embodiments, such as Figure 4 As shown, the median temperature, background average temperature, and segmental asymmetry index of the micro-thermal unit are fused to obtain the fusion coefficient, which includes:

[0079] S531. Calculate the absolute value of the difference between the median temperature and the background average temperature value to obtain the absolute temperature difference. , This represents the magnitude of local temperature changes.

[0080] S532. Calculate the product of the segmental asymmetry index and the absolute temperature difference to obtain the asymmetric temperature difference coupling quantity. .

[0081] The value couples the overall segmental asymmetry with the local temperature difference; if the segment to which the point belongs has high asymmetry, i.e. The value is relatively large, and the local temperature difference at that point is also large, that is... If the value is large, then The value will be very large, suggesting that this may be a hotspot strongly correlated with segmental dysfunction.

[0082] S533. Calculate the coupling strength based on the asymmetric temperature difference coupling amount and the absolute temperature difference. , This represents the coupling strength.

[0083] denominator Used for coupling quantities of asymmetric temperature difference Normalization is performed when the absolute temperature difference When it is very large, even if the segmental asymmetry index is large Generally, product It may also be larger, divided by This can reduce the influence of absolute temperature difference, making It focuses more on measuring the strength of the asymmetry coupled by a unit temperature difference.

[0084] S534. Calculate the fusion coefficient based on the asymmetric temperature difference coupling amount and coupling strength. .

[0085] The fusion coefficient C is constructed as the baseline value. Plus and fusion coefficient It can respond to the absolute magnitude of coupling It can also respond to the relative efficiency of coupling. ,when and When both are relatively large, It is significantly greater than 1, thus strongly enhancing the value of the local thermal gradient G at that point.

[0086] In some embodiments, determining whether a preset condition is met based on the segmental asymmetry index, the local thermal gradient value, and the hotspot concentration includes:

[0087] Determine whether the segment asymmetry index is greater than a preset first threshold, whether the local thermal gradient value is greater than a preset second threshold, and whether the hotspot concentration is greater than a preset third threshold.

[0088] The segmental asymmetry index of a large number of people without lumbar symptoms was statistically analyzed, and its distribution was calculated. The first threshold was set as a specific high quantile of the distribution, such as the 95th quantile, to automatically identify abnormal asymmetries that deviate significantly from the normal range.

[0089] The distribution of local thermal gradient values ​​of a large number of people without lumbar symptoms was calculated by statistical analysis. The second threshold was set as a specific high quantile of the distribution, such as the 95th quantile, to automatically screen out local thermal protrusions that are significantly higher than the normal background fluctuation.

[0090] The distribution of hotspot clusters in a large number of people without lumbar symptoms was statistically analyzed. The third threshold was set as an empirical quantile, such as the 70th percentile, that can distinguish between diffuse temperature rise and localized hotspots, so as to automatically identify areas with abnormally concentrated high-temperature pixels.

[0091] For example, the three final characteristic values ​​of the right facet joint of L4 in the subject: segmental asymmetry index Local thermal gradient value Hotspot concentration Assume the system's preset first threshold, second threshold, and third threshold are respectively... , , .because , ,and Since all three conditions were met, the system ultimately determined that the right facet joint region of the subject's L4 was an abnormal region.

[0092] In some embodiments, this application provides a system for identifying abnormal thermal distribution areas in waist infrared thermal images, including:

[0093] Landmark Recognition Module: Used to acquire mid- and far-infrared thermal images of the posterior lumbar region of the subject and identify key anatomical landmarks on the mid- and far-infrared thermal images.

[0094] Coordinate mapping module: used to identify the projection points of the bilateral facet joints of each lumbar vertebral segment in mid- and far-infrared thermal images based on key anatomical landmarks on the body surface.

[0095] Feature extraction module: used to define micro-thermal units with the center of the projection point, extract the statistical features of temperature distribution of micro-thermal units, generate a standard thermal matrix indexed by lumbar segments, and calculate the hot spot concentration of micro-thermal units.

[0096] Symmetry Analysis Module: Used to perform symmetry analysis on the temperature distribution statistical characteristics of the standard thermodynamic matrix, and generate a segmental asymmetry index that reflects the thermal differences between the two small joints.

[0097] Gradient Calculation Module: Used to calculate the thermal gradient of each small joint's micro-thermal unit relative to its local background based on the standard thermal matrix and segmental asymmetry index, generating local thermal gradient values.

[0098] Anomaly Detection Module: For each small joint's projection point, based on the segment asymmetry index, local thermal gradient value, and hotspot concentration, it determines whether preset conditions are met. If so, the micro-thermal unit defined by the small joint's projection point is identified as an abnormal region.

[0099] In some embodiments, this application provides a readable storage medium, including: computer program instructions stored in the readable storage medium, wherein the computer program instructions are read and executed by a processor to perform the steps of a method for identifying abnormal thermal distribution areas in a waist infrared thermal image.

[0100] It should be noted that, in this application, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0101] Any references to memory, storage, database, or other media used in the embodiments provided in this application may include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory may include random access memory (RAM) or external cache memory.

[0102] Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.

Claims

1. A method for identifying abnormal thermal distribution areas in waist infrared thermal images, characterized in that, include: Acquire mid- and far-infrared thermal images of the posterior lumbar region of the subject and identify key anatomical landmarks on the mid- and far-infrared thermal images; Based on the key anatomical landmarks, the projection points of the bilateral facet joints of each lumbar vertebral segment are identified in the mid- and far-infrared thermal images. A micro-thermal unit is defined with the center of the projection point, the temperature distribution statistical characteristics of the micro-thermal unit are extracted, a standard thermal matrix indexed by lumbar vertebral segments is generated, and the hot spot concentration degree of the micro-thermal unit is calculated. A symmetry analysis is performed on the temperature distribution statistical characteristics of the standard thermodynamic matrix to generate a segmental asymmetry index that reflects the thermal differences between the two small joints. Based on the standard thermodynamic matrix and segmental asymmetry index, the thermodynamic gradient of each small joint micro-thermal unit relative to its local background is calculated, and a local thermodynamic gradient value is generated. For each small joint's projection point, the system determines whether the preset conditions are met based on the segment asymmetry index, local thermal gradient value, and hotspot aggregation degree. If so, the micro-thermal unit defined by the small joint's projection point is identified as an abnormal region.

2. The method based on claim 1, characterized in that, Based on the aforementioned key anatomical landmarks, the spinous processes of each lumbar vertebral segment and the projection points of the bilateral facet joints are identified in mid- and far-infrared thermal imaging, including: Identify the interspinous space between the fourth and fifth lumbar vertebrae; Using the interspinous space as a reference, the projection points of the bilateral facet joints of the remaining lumbar vertebral segments are calculated according to the preset vertebral height intervals and lateral distances.

3. The method based on claim 1, characterized in that, A micro-thermal unit is defined at the center of the projection point. The statistical characteristics of the temperature distribution of the micro-thermal unit are extracted, and a standard thermal matrix indexed by lumbar vertebral segments is generated, including: Centered on each of the projection points, a rectangular area of ​​fixed pixel size is extracted from the mid- and far-infrared thermal image as a micro thermal unit. Extract the temperature values ​​of all pixels within each micro-thermal unit, and calculate the median temperature of that micro-thermal unit; A standard thermal matrix is ​​generated by arranging the median temperatures of all micro-thermal units according to their respective lumbar vertebral segments.

4. The method based on claim 3, characterized in that, Calculating the hotspot concentration of a micro-thermal unit includes: For each micro-thermal unit, count the number of pixels whose internal temperature value is higher than the median temperature; The hotspot concentration is obtained by calculating the ratio of the number of pixels to the total number of pixels in the micro-thermal unit.

5. The method based on claim 1, characterized in that, A symmetry analysis is performed on the temperature distribution statistical characteristics of the standard thermodynamic matrix to generate a segmental asymmetry index reflecting the thermal differences between the two facet joints, including: For each lumbar vertebral segment, the median temperature of the micro-thermal units of the facet joints on both sides of the lumbar vertebral segment is extracted from the standard thermal matrix to obtain the median temperature on the left and right sides. Calculate the absolute difference between the median temperature on the left and the median temperature on the right; Based on the median temperature on the left side, the median temperature on the right side, and the absolute difference, a significance assessment of the difference is performed to obtain the segmental asymmetry index.

6. The method based on claim 5, characterized in that, Based on the median temperature on the left side, the median temperature on the right side, and the absolute difference, a significance assessment of the difference is performed to obtain the segmental asymmetry index, including: Divide the larger of the left and right median temperatures by the smaller of the smaller values ​​to obtain the thermal background unevenness of the lumbar vertebral segment; calculate the ratio of the absolute difference to the sum of the left and right median temperatures to obtain the standardized relative difference of the lumbar vertebral segment. Based on the aforementioned thermal background inequilibrium and standardized relative differences, the modulation factor is calculated. , Represents the modulation factor. Representing relative differences in standardization, Represents background imbalance; Based on the standardized relative differences and modulation factors, the segmental asymmetry index is calculated. , This represents the segmental asymmetry index.

7. The method based on claim 1, characterized in that, Based on the standard thermodynamic matrix and segmental asymmetry index, the thermodynamic gradient of each small joint's micro-thermal unit relative to its local background is calculated, generating local thermodynamic gradient values, including: For each small joint projection point, in the mid- and far-infrared thermal image, based on the outer boundary of the micro-thermal unit of the projection point, a fixed-width annular region is extended outward. The set of pixels within the annular region that do not belong to any micro-thermal unit is defined as the local background region of the projection point. Calculate the average temperature of all pixels within the local background region to obtain the average background temperature value; The median temperature, background average temperature, and segmental asymmetry index of the micro-thermal unit are fused to obtain the fusion coefficient. Based on the fusion coefficient, median temperature, and background average temperature, the local thermal gradient value of the small joint is calculated. , , , and These represent the local thermal gradient, fusion coefficient, median temperature, and background average temperature of the micro-thermal unit, respectively.

8. The method based on claim 7, characterized in that, Feature fusion was performed on the median temperature, background average temperature, and segmental asymmetry index of the micro-thermal unit to obtain the fusion coefficient, including: The absolute temperature difference is obtained by calculating the absolute value of the difference between the median temperature and the background average temperature. The product of the segment asymmetry index and the absolute temperature difference is calculated to obtain the asymmetric temperature difference coupling amount; Calculate the coupling strength based on the asymmetric temperature difference coupling amount and the absolute temperature difference. , Represents coupling strength. Represents the asymmetric temperature difference coupling quantity. Represents absolute temperature difference; Based on the asymmetric temperature difference coupling amount and coupling strength, the fusion coefficient is calculated. .

9. The method based on claim 1, characterized in that, Based on the segmental asymmetry index, local thermal gradient value, and hotspot concentration, determine whether the preset conditions are met, including: Determine whether the segment asymmetry index is greater than a preset first threshold, whether the local thermal gradient value is greater than a preset second threshold, and whether the hotspot concentration is greater than a preset third threshold.

10. A system for identifying abnormal thermal distribution areas in waist infrared thermal images, characterized in that, include: Landmark Recognition Module: Used to acquire mid- and far-infrared thermal images of the posterior lumbar region of the subject and identify key anatomical landmarks on the mid- and far-infrared thermal images; Coordinate mapping module: used to identify the projection points of the bilateral facet joints of each lumbar vertebral segment in the mid- and far-infrared thermal image based on the key anatomical landmarks on the body surface; Feature extraction module: used to define micro-thermal units with the center of the projection point, extract the temperature distribution statistical features of the micro-thermal units, generate a standard thermal matrix indexed by lumbar vertebral segments, and calculate the hot spot concentration degree of the micro-thermal units; Symmetry Analysis Module: Used to perform symmetry analysis on the temperature distribution statistical characteristics of the standard thermodynamic matrix and generate a segmental asymmetry index that reflects the thermal differences between the two small joints; Gradient calculation module: used to calculate the thermal gradient of each small joint's micro-thermal unit relative to its local background based on the standard thermal matrix and segmental asymmetry index, and generate local thermal gradient values; Anomaly Detection Module: For each small joint's projection point, based on the segment asymmetry index, local thermal gradient value, and hotspot concentration, it determines whether preset conditions are met. If so, the micro-thermal unit defined by the small joint's projection point is identified as an abnormal region.

Images