Intelligent recognition system for spinal images

Abstract

This invention relates to the field of intelligent spinal detection technology, and more particularly to an intelligent recognition system for spinal images. The invention acquires the spinal contour through an image acquisition module, divides the contour into several sub-regions through a scanning module, and acquires the grayscale values ​​of several pixels within each sub-region. A grayscale calibration unit determines the calibrated grayscale values ​​for filling each pixel within the sub-regions. A filling unit determines the intervertebral space filling characterization and identifies the dominant sub-regions of intervertebral space features. A recognition module identifies the end-vertebral tangent within the dominant sub-regions of intervertebral space features. An evaluation output module outputs the Cobb angle of the scoliosis assessment result. Furthermore, this invention achieves precise analysis of pixel grayscale values ​​within the spinal contour sub-regions and explicit processing of intervertebral space features, establishing an effective mechanism for locating end-vertebral tangents based on pixel grayscale differences, thereby improving the accuracy and efficiency of Cobb angle calculation in scoliosis assessment.

Description

Technical Field

[0001] This invention relates to the field of intelligent spinal detection technology, and more particularly to an intelligent recognition system for spinal images. Background Technology

[0002] Scoliosis, a common three-dimensional spinal deformity, has a high incidence rate, especially among adolescents. The core of its diagnosis and treatment lies in accurately assessing the degree of curvature based on the Cobb angle to correct the curvature in a timely manner. Traditional Cobb angle measurements not only rely on manual operation by doctors, resulting in high subjectivity, large errors, low efficiency, and a high dependence on experience, but also suffer from blurred intervertebral space features in spinal images, increasing the difficulty of assessment. Traditional spinal image analysis techniques have bottlenecks in automated feature extraction, intervertebral space feature quantification, and the automation and accuracy of balanced Cobb angle calculation. Recent advancements in imaging technology have offered a solution to these challenges, leading to the development of intelligent spinal image recognition systems.

[0003] For example, Chinese Patent Publication No. CN120495377A discloses a method, apparatus, device, and storage medium for determining the Cobb angle of scoliosis based on Mamba. This method is applied to the field of scoliosis prediction. The method involves acquiring image information and preprocessing it to obtain an original image; extracting multi-level local features from the original image using an encoder to obtain a first feature map; capturing global features from the first feature map using a Mamba module to obtain a second feature map; decoding the second feature map using a decoder to obtain a target feature map; generating a center heatmap, offset map, and centripetal vector map based on the target feature map using a prediction module; extracting the coordinates of the center point of the vertebral body based on the center heatmap, offset map, and centripetal vector map; determining the coordinates of the four corner points of the vertebral body based on the coordinates of the center point; and determining the Cobb angle using a geometric algorithm based on the coordinates of the four corner points of the vertebral body.

[0004] The following problems still exist in the existing technology:

[0005] Existing technologies do not consider the precise analysis of pixel grayscale values ​​within the spinal contour sub-region and the explicit processing of intervertebral space features. They cannot establish an effective mechanism for locating the vertebral tangent by pixel grayscale differences, which affects the accuracy and efficiency of Cobb angle measurement in scoliosis assessment. Summary of the Invention

[0006] To address this issue, the present invention provides an intelligent recognition system for spinal images, which overcomes the problem that existing technologies cannot establish an effective mechanism for locating the vertebral tangent at the end of the spine due to pixel grayscale differences, thus affecting the accuracy and efficiency of Cobb angle calculation in scoliosis assessment.

[0007] To achieve the above objectives, the present invention provides an intelligent recognition system for spinal images, comprising:

[0008] An image acquisition module is used to acquire a spinal image of a target object and to acquire a spinal contour based on the spinal image.

[0009] The scanning module, which is connected to the image acquisition module, is used to divide the spinal contour into several contour sub-regions and acquire the gray values ​​of several pixels in each contour sub-region.

[0010] The intelligent estimation module, which is connected to the scanning module, includes a grayscale calibration unit and a filling unit. The grayscale calibration unit is used to determine the calibration grayscale value for filling each pixel in the contour sub-region based on the grayscale values ​​of several pixels in the contour sub-region.

[0011] The filling unit is used to fill each pixel in each contour sub-region, and to determine the intervertebral space filling characterization quantity based on the change of the gray value of the pixel in the contour sub-region before and after filling, so as to determine the dominant sub-region of intervertebral space features.

[0012] The recognition module, which is connected to the intelligent estimation module, is used to obtain the difference in gray values ​​of pixels that meet the preset constraints within the visible sub-region of the intervertebral space feature before filling, so as to determine the end vertebra tangent within the visible sub-region of the intervertebral space feature.

[0013] An evaluation output module, which is connected to the recognition module, is used to determine several Cobb angles by taking the end vertebral tangents in any two intervertebral disc feature dominant sub-regions as a reference, and outputting the largest Cobb angle as the scoliosis evaluation result.

[0014] Furthermore, the scanning module is used to establish a Cartesian coordinate system within the contour sub-region to determine the coordinates of each pixel in the Cartesian coordinate system, and to establish a correspondence between the coordinates of each pixel and the grayscale value. The vertical coordinate of the Cartesian coordinate system is set along the length of the spine.

[0015] Further, the grayscale calibration unit is used to determine the calibration grayscale value for filling each pixel point within the contour sub-region, wherein,

[0016] The grayscale calibration unit is used to pre-obtain the grayscale value of each pixel within the contour sub-region, and determine the average grayscale value of all pixels as the calibration grayscale value.

[0017] Furthermore, the filling unit is used to determine characteristic display pixels, including:

[0018] The filling unit is used to determine the gray value corresponding to the pixel before filling as the initial gray value, and to determine the absolute value of the difference between the calibrated gray value and the initial gray value as the gray value change amount;

[0019] The filling unit is used to identify pixels whose grayscale variation is greater than a preset grayscale variation threshold as feature display pixels.

[0020] Furthermore, the filling unit is used to define a dominant sub-region of intervertebral disc features, wherein,

[0021] The filling unit is used to determine the pixel area ratio of characteristic display pixels in the outline sub-region;

[0022] If the pixel area ratio is greater than a preset pixel area ratio threshold, the filling unit determines the contour sub-region as a prominent sub-region of intervertebral space features.

[0023] Furthermore, the recognition module determines pixels with the same ordinate value as pixels that satisfy preset constraints based on the coordinates of each pixel.

[0024] Furthermore, the recognition module is used to determine the difference in grayscale values ​​of pixels that meet preset constraints before filling, wherein,

[0025] The recognition module is used to divide the pixels that meet the preset constraints into a first set of pixels and a second set of pixels based on the region segmentation line of the prominent sub-region of the intervertebral space feature, and to obtain the first gray average value of the pixels in the first set of pixels and the second gray average value of the pixels in the second set of pixels before filling.

[0026] The recognition module determines the absolute value of the difference between the first grayscale average value and the second grayscale average value as the grayscale difference quantity.

[0027] Furthermore, the region segmentation line is a segmentation midline determined based on the abscissa of each pixel within the dominant subregion of the intervertebral space feature.

[0028] Furthermore, the identification module is used to determine the tangent reference point, wherein,

[0029] The recognition module is used to determine the first set of pixels and the second set of pixels corresponding to the maximum value of grayscale difference, and to determine the ordinate of the pixels in the first set of pixels and the second set of pixels as the ordinate of the tangent reference point, and to determine the abscissa of the region segmentation line as the abscissa of the tangent reference point.

[0030] Furthermore, the identification module is used to determine the tangent of the end vertebra, wherein,

[0031] The identification module identifies the line that passes through the tangent reference point and is parallel to the edge surface of the reference end vertebra as the end vertebra tangent;

[0032] The reference end-vertebra edge surface is the end-vertebra edge surface closest to the tangential reference point.

[0033] Compared with existing technologies, the beneficial effects of this invention are as follows: This invention sets up an image acquisition module, a scanning module, an intelligent estimation module, a recognition module, and an evaluation output module. The image acquisition module acquires the spinal contour; the scanning module divides the contour into several sub-regions and acquires the grayscale values ​​of several pixels within each sub-region; a grayscale calibration unit determines the calibrated grayscale values ​​for filling each pixel within the contour sub-regions; a filling unit determines the intervertebral space filling characterization and identifies the dominant sub-regions of intervertebral space features; the recognition module identifies the end-vertebral tangents within the dominant sub-regions of intervertebral space features; and the evaluation output module outputs the Cobb angle of the scoliosis evaluation result. Furthermore, this achieves accurate analysis of the grayscale values ​​of pixels within the spinal contour sub-regions and explicit processing of intervertebral space features, establishing an effective mechanism for locating end-vertebral tangents based on pixel grayscale differences, thus improving the accuracy and efficiency of Cobb angle calculation in scoliosis evaluation.

[0034] Furthermore, this invention uses a grayscale calibration unit to determine the average grayscale value of all pixels within the contour sub-region as the calibration grayscale value. This is significant in accurately distinguishing between the vertebral body and intervertebral space regions within the spinal contour. As is well known to those skilled in the art, the spinal contour sub-region includes two types of structures: the vertebral body and the intervertebral space. These two structures differ in grayscale values. The vertebral body region, due to its high bone density, has relatively concentrated and stable pixel grayscale values; the intervertebral space region, containing soft tissue or voids, exhibits grayscale values ​​that deviate from those of the vertebral body region. By calculating the average grayscale value of all pixels as the calibration grayscale value, a reference standard reflecting the dominant grayscale characteristics within the region is constructed—the grayscale level dominated by the vertebral body region. When this calibration grayscale value is used to fill the entire sub-region, the grayscale values ​​of pixels in the vertebral body region are close to the average value, with minimal changes before and after filling; while the grayscale values ​​of pixels in the intervertebral space region deviate significantly from the average value, resulting in substantial changes before and after filling. This ensures the representativeness of the calibration grayscale value for the regional features and improves adaptability under different imaging conditions.

[0035] Furthermore, this invention precisely distinguishes between vertebral bodies and intervertebral spaces within the spinal contour sub-region by quantifying and filtering the grayscale changes of pixels before and after filling. It can be understood that the filling unit uses the absolute value of the difference between the initial grayscale value and the calibrated grayscale value as the grayscale change. Utilizing the characteristic that pixels in the vertebral body region have small changes due to their initial grayscale being close to the calibrated grayscale, while pixels in the intervertebral space region have large changes due to their initial grayscale deviating from the calibrated grayscale, and then filtering out pixels with changes exceeding the threshold through a preset threshold, the pixels corresponding to the intervertebral spaces can be clearly separated. This solves the problem of blurred and difficult-to-distinguish grayscale between the intervertebral spaces and vertebral bodies caused by imaging interference in traditional technologies, achieving precise analysis of the grayscale values ​​of pixels within the spinal contour sub-region.

[0036] Furthermore, this invention provides a quantifiable basis for accurately identifying the morphological features of scoliosis by quantifying the grayscale difference between the left and right sides of pixels at the same height within the dominant subregion of the intervertebral space feature. The identification module divides the pixels at the same height into two groups based on the region segmentation line. By calculating the difference in the average grayscale value of the two groups of pixels before filling, the morphological difference of the intervertebral space in the left and right directions can be objectively reflected. This difference is precisely the manifestation of the intervertebral space widening on one side and narrowing on the other side during scoliosis, accurately reflecting the morphological imbalance of the intervertebral space caused by scoliosis.

[0037] Furthermore, specifically, the identification module of the present invention identifies the line that passes through the tangent reference point and is parallel to the nearest end vertebral edge surface as the end vertebral tangent. The reference end vertebral edge surface is the vertebral edge closest to the reference point, and its direction directly reflects the tilt state of the vertebral body in that segment. The tangent passes through the reference point at the most significant position of scoliosis and is parallel to the edge surface, ensuring that the tangent can accurately capture the scoliosis angle characteristics of that segment. This determination method solves the subjectivity problem when manually drawing tangents, establishes an effective mechanism for locating end vertebral tangents based on pixel grayscale differences, and improves the accuracy and efficiency of Cobb angle calculation in scoliosis assessment. Attached Figure Description

[0038] Figure 1 This is a block diagram of an intelligent recognition system for spinal images according to an embodiment of the present invention;

[0039] Figure 2 This is a flowchart illustrating the logic for determining characteristic display pixels in an embodiment of the present invention.

[0040] Figure 3 A flowchart illustrating the logic for determining the dominant sub-regions of intervertebral disc features in an embodiment of the present invention;

[0041] Figure 4 This is a schematic diagram illustrating the determination of the tangent line of the end vertebra in an embodiment of the present invention;

[0042] In the diagram: 1-Tangent reference point, 2-Reference end vertebral edge surface, 3-End vertebral tangent, 4-Vertebral bone. Detailed Implementation

[0043] To make the objectives and advantages of the present invention clearer, the present invention will be further described below with reference to embodiments; it should be understood that the specific embodiments described herein are merely for explaining the present invention and are not intended to limit the present invention.

[0044] Preferred embodiments of the present invention will now be described with reference to the accompanying drawings. Those skilled in the art should understand that these embodiments are merely illustrative of the technical principles of the present invention and are not intended to limit the scope of protection of the present invention.

[0045] It should be noted that in the description of this invention, the terms "upper," "lower," "inner," "outer," etc., which indicate the direction or positional relationship, are based on the direction or positional relationship shown in the drawings. This is only for the convenience of description and is not intended to indicate or imply that the device or element must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, it should not be construed as a limitation of this invention.

[0046] Furthermore, it should be noted that, in the description of this invention, unless otherwise explicitly specified and limited, the terms "installation" and "connection" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art can understand the specific meaning of the above terms in this invention according to the specific circumstances.

[0047] Please see Figure 1 The diagram shown is a block diagram of an intelligent recognition system for spinal images according to an embodiment of the present invention. The intelligent recognition system for spinal images of the present invention includes:

[0048] An image acquisition module is used to acquire a spinal image of a target object and to acquire a spinal contour based on the spinal image.

[0049] Specifically, the image acquisition module is not limited in this invention. It includes a CT imaging instrument for acquiring spinal images and an image processor connected to the CT imaging instrument to extract the spinal contour from the spinal images according to the edge contour algorithm. The CT imaging instrument in this invention has a resolution of 512×512-1024×1024 and a scanning slice thickness of 1-3mm for spinal images, which will not be described in detail here.

[0050] The scanning module, which is connected to the image acquisition module, is used to divide the spinal contour into several contour sub-regions and acquire the gray values ​​of several pixels in each contour sub-region.

[0051] In this invention, the spinal contour is divided into several contour sub-regions along the length of the spine. Each contour sub-region contains the contour lines on both sides of the spinal contour, and each contour sub-region contains 5 vertebrae.

[0052] Specifically, the present invention does not limit the scanning module, which may include a processor for dividing the spine contour, a grayscale image sensor for acquiring grayscale values ​​of several pixels, a memory for outputting pixel grayscale data using the RGB algorithm to collect grayscale signals, and a memory for storing the grayscale data.

[0053] The intelligent estimation module, which is connected to the scanning module, includes a grayscale calibration unit and a filling unit. The grayscale calibration unit is used to determine the calibration grayscale value for filling each pixel in the contour sub-region based on the grayscale values ​​of several pixels in the contour sub-region.

[0054] Specifically, the grayscale calibration unit in this invention can be a data processor, used to receive and calculate the grayscale values ​​of several pixels, which will not be elaborated here.

[0055] The filling unit is used to fill each pixel in each contour sub-region, and to determine the intervertebral space filling characterization quantity based on the change of the gray value of the pixel in the contour sub-region before and after filling, so as to determine the dominant sub-region of intervertebral space features.

[0056] Specifically, the filling unit in this invention can be a writing device that implements the writing of the target filling value and a register that calculates the absolute value of the difference between different data groups.

[0057] The recognition module, which is connected to the intelligent estimation module, is used to obtain the difference in gray values ​​of pixels that meet the preset constraints within the visible sub-region of the intervertebral space feature before filling, so as to determine the end vertebra tangent within the visible sub-region of the intervertebral space feature.

[0058] Specifically, the identification module in this invention can be a processor with data processing and logic control capabilities, which will not be elaborated here.

[0059] An evaluation output module, which is connected to the recognition module, is used to determine several Cobb angles by taking the end vertebral tangents in any two intervertebral disc feature dominant sub-regions as a reference, and outputting the largest Cobb angle as the scoliosis evaluation result.

[0060] In this invention, determining the Cobb angle based on the angle formed by the perpendicular lines of the two end vertebral tangents is a common technique used by those skilled in the art in the process of identifying scoliosis, and will not be elaborated here.

[0061] Specifically, the evaluation output module in this invention can be a computer or a micro-control machine, which can read program code pre-stored in a storage medium to perform the task of outputting scoliosis evaluation results.

[0062] Specifically, this invention improves the accuracy and efficiency of Cobb angle calculation in scoliosis assessment. The accuracy of the Cobb angle directly depends on the positioning accuracy of the end vertebral tangents. This invention uses grayscale difference analysis to locate the reference point and finally determines the end vertebral tangents, ensuring that each end vertebral tangent corresponds to the most significant position of scoliosis. The Cobb angle calculated by the assessment output module based on these end vertebral tangents can truly reflect the severity of scoliosis. The determination of the maximum Cobb angle provides accurate quantitative results.

[0063] Specifically, the scanning module is used to establish a Cartesian coordinate system within the contour sub-region to determine the coordinates of each pixel in the Cartesian coordinate system and to establish a correspondence between the coordinates of each pixel and the grayscale value. The vertical coordinate of the Cartesian coordinate system is set along the length of the spine.

[0064] Specifically, the grayscale calibration unit is used to determine the calibration grayscale value for filling each pixel point within the contour sub-region, wherein,

[0065] The grayscale calibration unit is used to pre-obtain the grayscale value of each pixel within the contour sub-region, and determine the average grayscale value of all pixels as the calibration grayscale value.

[0066] In this invention, for each pixel within the contour sub-region, the corresponding red, green, and blue channel values ​​are read from the image data, and the grayscale value is calculated using the RGB to grayscale formula: grayscale value = 0.299×R + 0.587×G + 0.114×B. This process will not be elaborated further here.

[0067] It is understood that this invention uses a grayscale calibration unit to determine the average grayscale value of all pixels within a contour sub-region as the calibration grayscale value. This is significant in accurately distinguishing between the vertebral bodies and intervertebral spaces within the spinal contour. As is well known to those skilled in the art, the spinal contour sub-region includes two types of structures: the vertebral body and the intervertebral space. These two structures differ in grayscale values. The vertebral body region, due to its high bone density, has relatively concentrated and stable pixel grayscale values; the intervertebral space region, containing soft tissue or voids, exhibits grayscale values ​​that deviate from those of the vertebral body region. By calculating the average grayscale value of all pixels as the calibration grayscale value, a reference standard reflecting the dominant grayscale characteristics within the region is constructed—the grayscale level dominated by the vertebral body region. When this calibration grayscale value is used to fill the entire sub-region, the grayscale values ​​of pixels in the vertebral body region are close to the average value, with minimal changes before and after filling; while the grayscale values ​​of pixels in the intervertebral space region deviate significantly from the average value, resulting in substantial changes before and after filling. This ensures the representativeness of the calibration grayscale value for the regional features and improves adaptability under different imaging conditions.

[0068] Specifically, please refer to Figure 2 The diagram shown is a flowchart illustrating the logic for determining feature display pixels according to an embodiment of the present invention. The filling unit is used to determine feature display pixels and includes:

[0069] The filling unit is used to determine the gray value corresponding to the pixel before filling as the initial gray value, and to determine the absolute value of the difference between the calibrated gray value and the initial gray value as the gray value change amount;

[0070] The filling unit is used to identify pixels whose grayscale variation is greater than a preset grayscale variation threshold as feature display pixels.

[0071] The filling unit does not determine the characteristic display pixels for pixels whose grayscale change is less than or equal to a preset grayscale change threshold.

[0072] In this invention, the value of the preset grayscale change threshold can be determined based on the results of prior experimental statistics. The average grayscale change before and after pixel filling in the vertebral region of the spinal image obtained from prior experimental statistics is determined as the grayscale change threshold. Here, a value of 20 is provided for the grayscale change threshold.

[0073] Specifically, this invention precisely distinguishes between vertebral bodies and intervertebral spaces within the spinal contour sub-region by quantifying and filtering the grayscale changes of pixels before and after filling. It can be understood that the filling unit uses the absolute value of the difference between the initial grayscale value and the calibrated grayscale value as the grayscale change. Utilizing the characteristic that pixels in the vertebral body region have small changes due to their initial grayscale being close to the calibrated grayscale, while pixels in the intervertebral space region have large changes due to their initial grayscale deviating from the calibrated grayscale, a preset threshold is used to filter out pixels with changes exceeding the threshold. This clearly separates the pixels corresponding to the intervertebral spaces, solving the problem of blurred and difficult-to-distinguish grayscale between the intervertebral spaces and vertebral bodies caused by imaging interference in traditional technologies, and achieving precise analysis of the grayscale values ​​of pixels within the spinal contour sub-region.

[0074] Specifically, please refer to Figure 3 As shown, this is a flowchart illustrating the logic for determining the dominant sub-region of intervertebral disc features according to an embodiment of the present invention. The filling unit is used to determine the dominant sub-region of intervertebral disc features, wherein...

[0075] The filling unit is used to determine the pixel area ratio of characteristic display pixels in the outline sub-region;

[0076] If the pixel area ratio is greater than a preset pixel area ratio threshold, the filling unit determines the contour sub-region as a dominant sub-region of intervertebral space features.

[0077] If the pixel area ratio is less than or equal to a preset pixel area ratio threshold, the filling unit will not determine the contour sub-region as a prominent sub-region of intervertebral disc features.

[0078] In this invention, to avoid misjudgment due to an excessively small pixel area percentage threshold and missed judgment due to an excessively large pixel area percentage threshold, those skilled in the art can set the range of the pixel area percentage threshold to [8%, 13%]. Preferably, the pixel area percentage threshold is 10%.

[0079] It is understood that this invention accurately filters out regions with significant intervertebral disc features by comparing the area ratio of characteristic display pixels with a preset threshold. The characteristic display pixels centrally reflect the grayscale change characteristics of the intervertebral disc region. The outline sub-region whose area ratio exceeds the threshold means that the intervertebral disc features in that region are sufficiently obvious. Those skilled in the art understand that when the spine tends to bend laterally due to factors such as asymmetrical muscle strength or abnormal vertebral development, the forces on the two vertebrae show significant differences: on the convex side, the lateral bending pull causes the intervertebral discs to be stretched in the convex direction, thereby increasing the intervertebral disc in that direction and making the intervertebral disc more prominent. This achieves accurate analysis of the grayscale values ​​of pixels within the spinal outline sub-region and explicit processing of intervertebral disc features.

[0080] Specifically, the recognition module determines pixels with the same ordinate value as pixels that meet preset constraints based on the coordinates of each pixel.

[0081] Specifically, the recognition module in this invention defines pixels with the same ordinate value as pixels that meet preset constraints. In essence, it determines the analysis dimension of the same height position of the spine. In spinal imaging, whether the morphology of the intervertebral spaces on the left and right sides at the same height is symmetrical is an important indicator for judging whether scoliosis has occurred. Those skilled in the art understand that the gray distribution of the intervertebral spaces on the left and right sides at the same height of a normal spine is uniform, while the scoliosis spine has asymmetrical changes on the convex and concave sides, thus enabling the identification of the asymmetrical features of the intervertebral spaces.

[0082] Specifically, the recognition module is used to determine the difference in grayscale values ​​of pixels that meet preset constraints before filling, wherein,

[0083] The recognition module is used to divide the pixels that meet the preset constraints into a first set of pixels and a second set of pixels based on the region segmentation line of the prominent sub-region of the intervertebral space feature, and to obtain the first gray average value of the pixels in the first set of pixels and the second gray average value of the pixels in the second set of pixels before filling.

[0084] The recognition module determines the absolute value of the difference between the first grayscale average value and the second grayscale average value as the grayscale difference quantity.

[0085] Understandably, this invention provides a quantifiable basis for accurately identifying the morphological features of scoliosis by quantifying the grayscale difference between the left and right sides of pixels at the same height within the dominant subregion of the intervertebral space feature. The identification module divides pixels at the same height into two groups based on the region segmentation line. By calculating the difference in the average grayscale value of the two groups of pixels before filling, the morphological difference of the intervertebral space in the left and right directions can be objectively reflected. This difference is precisely the manifestation of the intervertebral space widening on one side and narrowing on the other side during scoliosis, accurately reflecting the morphological imbalance of the intervertebral space caused by scoliosis.

[0086] Specifically, the region segmentation line is a segmentation midline determined based on the abscissa of each pixel within the dominant subregion of the intervertebral space feature.

[0087] Specifically, the identification module is used to determine the tangent reference point, wherein,

[0088] The recognition module is used to determine the first set of pixels and the second set of pixels corresponding to the maximum value of grayscale difference, and to determine the ordinate of the pixels in the first set of pixels and the second set of pixels as the ordinate of the tangent reference point, and to determine the abscissa of the region segmentation line as the abscissa of the tangent reference point.

[0089] Specifically, the recognition module of this invention uses the ordinate of the set of pixels corresponding to the maximum grayscale difference as the ordinate of the tangent reference point. In essence, it determines the segment with the most significant scoliosis. The maximum grayscale difference means that the intervertebral space at that height is most asymmetrical, which is the position with the most prominent scoliosis feature in the scoliosis curve. At the same time, the abscissa of the region segmentation line is used as the abscissa of the reference point to ensure that the point is located at the midline of the intervertebral space, providing an anchor point for subsequent tangent drawing.

[0090] Specifically, the identification module is used to determine the tangent of the end vertebra, wherein,

[0091] The identification module identifies the line that passes through the tangent reference point 1 and is parallel to the reference end vertebra edge surface 2 as the end vertebra tangent 3;

[0092] The reference end vertebral edge surface 2 is the end vertebral edge surface closest to the tangent reference point 1.

[0093] Please see Figure 4 As shown, it is a schematic diagram of determining the vertebral tangent in an embodiment of the present invention. The contour sub-region contains 5 vertebrae 4. The reference vertebral edge surface 2 is the vertebral edge surface closest to the tangent reference point 1. The line that passes through the tangent reference point 1 and is parallel to the reference vertebral edge surface 2 is determined as the vertebral tangent 3.

[0094] In this invention, the edge surface of the end vertebra is the upper or lower edge surface of the end vertebra, and the distance between the tangent reference point and the edge surface of the end vertebra is a straight line distance.

[0095] In this invention, the upper or lower edge surface of the end vertebra is determined by an edge detection algorithm, which will not be elaborated here.

[0096] Specifically, the identification module of this invention identifies the line that passes through the tangent reference point and is parallel to the nearest end vertebral edge surface as the end vertebral tangent. The reference end vertebral edge surface is the vertebral body edge closest to the reference point, and its direction directly reflects the tilt state of the vertebral body in that segment. The tangent passes through the reference point at the most significant position of scoliosis and is parallel to the edge surface, ensuring that the tangent can accurately capture the scoliosis angle characteristics of that segment. This determination method solves the subjectivity problem when manually drawing tangents, establishes an effective mechanism for locating end vertebral tangents based on pixel grayscale differences, and improves the accuracy and efficiency of Cobb angle calculation in scoliosis assessment.

[0097] The technical solution of the present invention has been described above with reference to the preferred embodiments shown in the accompanying drawings. However, it will be readily understood by those skilled in the art that the scope of protection of the present invention is obviously not limited to these specific embodiments. Without departing from the principles of the present invention, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions after these changes or substitutions will all fall within the scope of protection of the present invention.

[0098] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. An intelligent recognition system for spinal images, characterized in that, include: An image acquisition module is used to acquire a spinal image of a target object and to acquire a spinal contour based on the spinal image. The scanning module, which is connected to the image acquisition module, is used to divide the spinal contour into several contour sub-regions and acquire the gray values ​​of several pixels in each contour sub-region. The intelligent estimation module, which is connected to the scanning module, includes a grayscale calibration unit and a filling unit. The grayscale calibration unit is used to determine the calibration grayscale value for filling each pixel in the contour sub-region based on the grayscale values ​​of several pixels in the contour sub-region. The filling unit is used to fill each pixel in each contour sub-region, and to determine the intervertebral space filling characterization quantity based on the change of the gray value of the pixel in the contour sub-region before and after filling, so as to determine the dominant sub-region of intervertebral space features. The filling unit is used to determine the gray value corresponding to the pixel before filling as the initial gray value, determine the absolute value of the difference between the calibrated gray value and the initial gray value as the gray value change amount, and determine the pixel with the gray value change amount greater than the preset gray value change amount threshold as the feature display pixel. The filling unit is used to determine the pixel area ratio of characteristic display pixels in the contour sub-region. If the pixel area ratio is greater than a preset pixel area ratio threshold, the filling unit determines the contour sub-region as a prominent sub-region of intervertebral disc features. The identification module, which is connected to the intelligent estimation module, is used to obtain the difference in gray values ​​of pixels that meet preset constraints before filling within the visible sub-region of the intervertebral disc feature. Based on the difference, the visible sub-region of the intervertebral disc feature with morphological imbalance is selected to determine the end vertebral tangent within the visible sub-region of the intervertebral disc feature. The recognition module identifies pixels with the same ordinate value as pixels that meet preset constraints based on the coordinates of each pixel. The recognition module is used to divide the pixels that meet the preset constraints into a first set of pixels and a second set of pixels based on the region segmentation line of the prominent sub-region of the intervertebral space feature, and to obtain the first gray value of the pixels in the first set of pixels and the second gray value of the pixels in the second set of pixels before filling, and to determine the absolute value of the difference between the first gray value and the second gray value as the gray value difference. The tangent of the end vertebra is a line that passes through the tangent reference point and is parallel to the edge surface of the reference end vertebra. The tangent reference point is determined based on the pixel coordinates in the first pixel set and the second pixel set. An evaluation output module, which is connected to the recognition module, is used to determine several Cobb angles by taking the end vertebral tangents in any two intervertebral disc feature dominant sub-regions as a reference, and outputting the largest Cobb angle as the scoliosis evaluation result.

2. The intelligent recognition system for spinal images according to claim 1, characterized in that, The scanning module is used to establish a Cartesian coordinate system within the contour sub-region to determine the coordinates of each pixel in the Cartesian coordinate system and to establish a correspondence between the coordinates of each pixel and the grayscale value. The vertical coordinate of the Cartesian coordinate system is set along the length of the spine.

3. The intelligent recognition system for spinal images according to claim 2, characterized in that, The grayscale calibration unit is used to determine the calibration grayscale value for filling each pixel point within the contour sub-region, wherein, The grayscale calibration unit is used to pre-obtain the grayscale value of each pixel within the contour sub-region, and determine the average grayscale value of all pixels as the calibration grayscale value.

4. The intelligent recognition system for spinal images according to claim 1, characterized in that, The region segmentation line is the segmentation midline determined based on the abscissa of each pixel within the dominant sub-region of the intervertebral space feature.

5. The intelligent recognition system for spinal images according to claim 4, characterized in that, The identification module is used to determine the tangent reference point, wherein, The recognition module is used to determine the first set of pixels and the second set of pixels corresponding to the maximum value of grayscale difference, and to determine the ordinate of the pixels in the first set of pixels and the second set of pixels as the ordinate of the tangent reference point, and to determine the abscissa of the region segmentation line as the abscissa of the tangent reference point.

6. The intelligent recognition system for spinal images according to claim 5, characterized in that, The identification module is used to determine the tangent of the end vertebra, wherein... The identification module identifies the line that passes through the tangent reference point and is parallel to the edge surface of the reference end vertebra as the end vertebra tangent; The reference end-vertebra edge surface is the end-vertebra edge surface closest to the tangential reference point.

Images