Mamba-based scoliosis cobb angle determination method, device, equipment and storage medium

Abstract

The embodiment of the application provides a kind of based on Mamba's scoliosis Cobb angle determination method, device, equipment and storage medium. Applied to scoliosis prediction field. The method is obtained by acquiring image information, and the original image is obtained by pre-processing the acquired image information;Through the encoder, the original image is extracted to the first feature map;Through Mamba module, the first feature map is captured to the second feature map;Through the decoder, the second feature map is decoded to obtain the target feature map;Through the prediction module, the center heat map, the offset map and the centripetal vector map are generated based on the target feature map;According to the center heat map, the offset map and the centripetal vector map, the center point coordinates of the vertebral body are extracted, and the four corner point coordinates of the vertebral body are determined according to the center point coordinates of the vertebral body;According to the four corner point coordinates of the vertebral body, the Cobb angle is determined by geometric algorithm, and the precision of Cobb angle measurement is improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of scoliosis prediction, and in particular to a Mamba-based scoliosis Cobb angle determination method, device, equipment and storage medium. BACKGROUND

[0002] Accurate assessment of scoliosis is crucial for clinical diagnosis and treatment decision-making. Traditional methods mainly rely on manual measurement of Cobb angle by doctors on X-ray images. However, manual measurement method is not only inefficient, time-consuming and laborious, but also easily affected by operator experience and subjective judgment, resulting in poor repeatability and low consistency, which is difficult to meet the needs of precision medicine.

[0003] To overcome the limitations of manual measurement method, automated measurement technology based on deep learning has emerged. For example, the Cobb angle measurement method based on convolutional neural network can automatically extract local features such as vertebral texture and edge from X-ray images for positioning key points or segmenting spine region, realizing automatic calculation of Cobb angle, thereby improving the efficiency and objectivity of evaluation.

[0004] However, existing automated measurement methods based on convolutional neural network or other single models are insufficient in capturing overall structural information and long-distance dependency of spine, which makes it difficult for the model to accurately understand the overall bending pattern, limiting the accuracy of Cobb angle measurement. SUMMARY

[0005] The present application provides a Mamba-based scoliosis Cobb angle determination method, device, equipment and storage medium to solve the problem that existing technology is insufficient in capturing overall structural information and long-distance dependency of spine, which limits the accuracy of Cobb angle measurement.

[0006] In a first aspect, the present application provides a Mamba-based scoliosis Cobb angle determination method, which comprises:

[0007] Obtaining image information and pre-processing the obtained image information to obtain an original image;

[0008] Performing multi-level local feature extraction on the original image through an encoder to obtain a first feature map, wherein the encoder comprises a plurality of residual modules;

[0009] Capturing global features of the first feature map through a Mamba module to obtain a second feature map;

[0010] Decoding the second feature map through a decoder to obtain a target feature map, wherein the decoder comprises a main decoding branch and an auxiliary decoding branch;

[0011] The center heat map, the offset map and the centripetal vector map are generated based on the target feature map by a prediction module;

[0012] The center point coordinates of the vertebral body are extracted according to the center heat map, the offset map and the centripetal vector map, and the four corner point coordinates of the vertebral body are determined according to the center point coordinates of the vertebral body;

[0013] The Cobb angle is determined by a geometric algorithm according to the four corner point coordinates of the vertebral body.

[0014] Optionally, the global feature capturing of the first feature map is performed by the Mamba module to obtain a second feature map, including:

[0015] The first feature map is subjected to feature serialization processing to obtain a first feature vector sequence;

[0016] The first feature vector sequence is analyzed and processed by a structured state space model of the Mamba module to obtain a second feature vector sequence;

[0017] The second feature vector sequence is mapped back to a two-dimensional feature map to obtain a second feature map.

[0018] Optionally, the second feature map is decoded by the decoder to obtain the target feature map, including:

[0019] The second feature map is subjected to up-sampling and convolution operation by the main decoding branch to obtain a third feature map;

[0020] The features of the intermediate layer of the encoder are received by the auxiliary decoding branch, and the auxiliary loss function is used for back propagation optimization to obtain a fourth feature map;

[0021] The third feature map and the fourth feature map are subjected to fusion processing to obtain the target feature map.

[0022] Optionally, the center point coordinates of the vertebral body are extracted according to the center heat map, the offset map and the centripetal vector map, and the four corner point coordinates of the vertebral body are determined according to the center point coordinates of the vertebral body, including:

[0023] The non-maximum suppression algorithm is applied to process the center heat map, and a preset number of peak points with the highest confidence score are selected in the processed center heat map, and for each selected peak point, the integer coordinates and the confidence score thereof on the target feature map are recorded;

[0024] The corresponding offset vector is extracted from the offset map according to the integer coordinates of the peak point, and the center point coordinates of the vertebral body are determined according to the integer coordinates and the offset vector;

[0025] The corresponding eight vector components are extracted from the centripetal vector map according to the center point coordinates of the vertebral body, and the four corner point coordinates of the vertebral body are determined according to the center point coordinates of the vertebral body and the extracted vector components.

[0026] Optionally, the Cobb angle is determined according to the four corner point coordinates of the vertebral body through a geometric algorithm, comprising:

[0027] According to the four corner point coordinates of the vertebral body, the direction line of each vertebral body is determined;

[0028] According to the direction line of the vertebral body, the angle between all pairs of vertebral bodies is determined to obtain an angle matrix;

[0029] The range of vertebral bodies corresponding to the main curve is determined, and the Cobb angle of the main curve is determined according to the angle matrix;

[0030] The range of vertebral bodies corresponding to the secondary curve is determined, and the Cobb angle of the secondary curve is determined through a geometric algorithm.

[0031] Optionally, the fourth feature map is obtained by back propagation optimization through an auxiliary loss function, comprising:

[0032] A first loss function is defined for the center heat map, and the first loss function is used to measure the difference between the predicted center heat map and the true center heat map;

[0033] A second loss function is defined for the offset map, and the second loss function is used to measure the difference between the predicted offset map and the true offset map;

[0034] A third loss function is defined for the centripetal vector map, and the third loss function is used to measure the difference between the predicted centripetal vector map and the true centripetal vector map;

[0035] The first loss function, the second loss function and the third loss function are weighted and summed to obtain a total loss;

[0036] The fourth feature map is obtained by back propagation optimization according to the total loss.

[0037] In a second aspect, the present application provides a Mamba-based scoliosis Cobb angle determination device, comprising:

[0038] An acquisition module is configured to acquire image information;

[0039] A first processing module is configured to preprocess the acquired image information to obtain an original image;

[0040] A second processing module is configured to extract multi-level local features of the original image through an encoder to obtain a first feature map, wherein the encoder comprises a plurality of residual modules;

[0041] The second processing module is further configured to capture global features of the first feature map through a Mamba module to obtain a second feature map;

[0042] The second processing module is further configured to decode the second feature map by a decoder to obtain a target feature map, wherein the decoder comprises a main decoding branch and an auxiliary decoding branch.

[0043] The second processing module is further configured to generate a center heat map, an offset map and a centripetal vector map based on the target feature map by a prediction module.

[0044] The second processing module is further configured to extract the center point coordinates of the vertebral body according to the center heat map, the offset map and the centripetal vector map, and determine the four corner point coordinates of the vertebral body according to the center point coordinates of the vertebral body.

[0045] The second processing module is further configured to determine the Cobb angle of the scoliosis according to the four corner point coordinates of the vertebral body by a geometric algorithm.

[0046] In a third aspect, the present application provides a Mamba-based scoliosis Cobb angle determination device, comprising:

[0047] a memory;

[0048] a processor;

[0049] The memory stores computer execution instructions.

[0050] The processor executes the computer execution instructions stored in the memory to implement the Mamba-based scoliosis Cobb angle determination method as described in the first aspect and various possible implementation manners of the first aspect.

[0051] In a fourth aspect, the present application provides a computer storage medium having a computer program stored thereon, wherein the computer program is executed by a processor to implement the Mamba-based scoliosis Cobb angle determination method as described in the first aspect and various possible implementation manners of the first aspect.

[0052] In a fifth aspect, the present application provides a computer program product comprising a computer program, wherein the computer program is executed by a processor to implement the Mamba-based scoliosis Cobb angle determination method as described in the first aspect and various possible implementation manners of the first aspect.

[0053] The application provides a Mamba-based scoliosis Cobb angle determination method, device, equipment and storage medium. The method obtains image information, and pre-processes the obtained image information to obtain an original image; performs multi-level local feature extraction on the original image through an encoder to obtain a first feature map; captures global features of the first feature map through a Mamba module to obtain a second feature map; decodes the second feature map through a decoder to obtain a target feature map; generates a center heat map, an offset map and a centripetal vector map based on the target feature map through a prediction module; extracts the center point coordinates of the vertebral body according to the center heat map, the offset map and the centripetal vector map, and determines the four corner point coordinates of the vertebral body according to the center point coordinates of the vertebral body; and determines the Cobb angle through a geometric algorithm according to the four corner point coordinates of the vertebral body, thereby improving the accuracy of Cobb angle measurement. BRIEF DESCRIPTION OF DRAWINGS

[0054] The accompanying drawings, which are incorporated herein and form part of the specification, illustrate embodiments consistent with the present application and, together with the description, further serve to explain the principles of the application.

[0055] Figure 1 A flowchart of a Mamba-based scoliosis Cobb angle determination method provided by an embodiment of the application is shown in the figure.

[0056] Figure 2 A structure diagram of a Mamba-based scoliosis Cobb angle determination device provided by an embodiment of the application is shown in the figure.

[0057] Figure 3 A structure diagram of a Mamba-based scoliosis Cobb angle determination device provided by an embodiment of the application is shown in the figure.

[0058] The above-described figures have shown the specific embodiments of the application, which will be described in more detail hereinafter. These figures and the written description are not intended to limit the scope of the concept of the application in any way, but to illustrate the concept of the application to those skilled in the art by referring to specific embodiments. DETAILED DESCRIPTION

[0059] To make the purposes, technical solutions and advantages of the application clearer, the technical solutions in the application will be described clearly and completely below by referring to the figures in the application. Obviously, the described embodiments are some of the embodiments of the application, but not all the embodiments. Based on the embodiments in the application, all other embodiments obtained by those skilled in the art without any creative work fall within the protection scope of the application.

[0060] The terms "first," "second," "third," "fourth," etc. (if present) in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein.

[0061] In this application, the terms "exemplary" or "for example" are used to indicate examples, illustrations, or descriptions. Any embodiment or design described as "exemplary" or "for example" in this application should not be construed as being more preferred or advantageous than other embodiments or designs. Specifically, the use of terms such as "exemplary" or "for example" is intended to present the relevant concepts in a specific manner.

[0062] The technical solution of this application and how the technical solution of this application solves the above-mentioned technical problems are described in detail below with specific embodiments. These specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described again in some embodiments. The embodiments of this application will now be described with reference to the accompanying drawings.

[0063] Figure 1 This is a flowchart illustrating a method for determining the Cobb angle of scoliosis based on Mamba, provided as an embodiment of this application. Figure 1 As shown, the Mamba-based method for determining the Cobb angle of scoliosis provided in this embodiment includes:

[0064] S1: Acquire image information and preprocess the acquired image information to obtain the original image.

[0065] The process of preprocessing the acquired image information to obtain the original image includes:

[0066] Data cropping: Remove areas below the brain, neck, and pelvis to ensure recognition stability;

[0067] Size adjustment: Adjust all input images to a fixed size (e.g., 1024×512) to ensure consistent data format;

[0068] Image enhancement: Images are enhanced through methods such as random cropping, expansion, contrast and brightness adjustment to increase data diversity and reduce the risk of overfitting;

[0069] Standardization: Normalize the pixel values, adjusting their range to [0,1] to facilitate model processing.

[0070] S2: The encoder performs multi-level local feature extraction on the original image to obtain the first feature map.

[0071] The encoder comprises a plurality of residual modules.

[0072] It can be understood that the original image enters the encoder part, and the initial convolutional layer of the convolutional neural network is used to extract multi-level local features in the original image. The extracted features contain key detail information such as texture, edge, and corner point of the vertebral body.

[0073] In an optional embodiment, ResNet34 is used as the backbone network, and the initial convolutional layer comprises Cov1, Cov2, Cov3, and Cov4. The original image is subjected to multi-level local feature extraction. Specifically, shallow convolution (Cov1 and Cov2) is used to capture low-level local features in the original image, such as edge contour and bone texture of the vertebral body, and the above features are helpful for accurate identification of the vertebral body boundary. With the deepening of the network, the receptive field gradually increases, and deep convolution (Cov3 and Cov4) can learn more complex and abstract combined features, such as corner point of the vertebral body and specific morphology of the vertebral plate.

[0074] It can also be understood that the residual connection mechanism of ResNet effectively alleviates the gradient vanishing problem in the training of deep networks, so that the network can learn deeper and more discriminative local features.

[0075] Further, after the processing of the backbone network, the original input image where H0 is the height of the original image, W0 is the width of the original image, and 3 is the number of channels, which is converted into a high-dimensional feature map F ∈ R (H×W×C) where, C = 512, and the output high-dimensional feature map contains rich and multi-scale local detail information of the vertebral body.

[0076] The purpose of this step is to use the local feature extraction capability of the convolutional neural network to provide high-quality and fine-grained local anatomical structure representation for subsequent global information modeling.

[0077] S3: capturing global features of the first feature map through the Mamba module to obtain a second feature map.

[0078] Specifically, the first feature map is subjected to feature serialization processing to obtain a first feature vector sequence; the first feature vector sequence is analyzed and processed through the structured state space model of the Mamba module to obtain a second feature vector sequence; and the second feature vector sequence is mapped back to a two-dimensional feature map to obtain a second feature map.

[0079] In an optional embodiment, the first feature map is subjected to feature serialization processing to obtain a first feature vector sequence, including: flattening the spatial dimension of HxW into a sequence with a length of HxW, and each sequence element is a C-dimensional feature vector; and traversing the blocks on the feature map using a specific scanning order (such as raster scanning, Z-order scanning) to form a feature vector sequence, which can preserve the original spatial proximity.

[0080] In an optional embodiment, the first feature vector sequence is analyzed and processed by a structured state space model of the Mamba module to obtain a second feature vector sequence, including: the structured state space model of the Mamba module compresses and transmits historical information through an implicit state h t The basic dynamics can be represented as:

[0081] x′ t = Bx t

[0082] h′ t = Ah t +x′ t

[0083] y t = Ch′ t

[0084] where x t is the current input feature vector in the sequence, x′ t is the intermediate feature vector, h′ is the updated hidden state, y t is the current output feature vector in the sequence, and A, B, and C are state space parameters.

[0085] It can be understood that the state space parameters can be dynamically adjusted according to the input, that is, the model can selectively “remember” or “forget” the information in the sequence according to what it has seen, thereby more effectively focusing on long-distance dependencies related to the current task. Mamba further enhances the model's ability to control information flow by combining a gating mechanism, enabling it to dynamically combine and filter information transmitted through the structured state space model.

[0086] The first feature map is subjected to global feature capture by the Mamba module to obtain a second feature map, which makes up for the deficiencies of traditional convolutional neural networks in global context understanding and long-distance dependency capture, and can effectively simulate the bending shape of the entire spine, the relative position and rotation relationship between the vertebral bodies, helping to accurately evaluate complex scoliosis patterns. Through the combination of convolutional neural networks and Mamba, effective fusion of local fine features and global structural information is achieved.

[0087] S4: decoding the second feature map through a decoder to obtain a target feature map.

[0088] The decoder includes a main decoding branch and an auxiliary decoding branch.

[0089] Specifically, the second feature map is up-sampled and convoluted through the main decoding branch to obtain a third feature map; the features of the intermediate layers of the encoder are received through the auxiliary decoding branch, and the fourth feature map is obtained through back propagation optimization of the auxiliary loss function; the third feature map and the fourth feature map are fused to obtain the target feature map.

[0090] It can be understood that the main decoding branch gradually recovers the spatial resolution through up-sampling and convolution operations to generate the final feature map; the auxiliary decoding branch uses a self-supervised learning strategy to supervise the intermediate features to improve the feature expression capability.

[0091] In an optional embodiment, the fourth feature map is obtained through back propagation optimization of the auxiliary loss function, including: defining a first loss function for the center heat map, the first loss function being used to measure the difference between the predicted center heat map and the real center heat map, the first loss function being, for example, a focal loss; defining a second loss function for the offset map, the second loss function being used to measure the difference between the predicted offset map and the real offset map, the second loss function being, for example, an L1 loss; defining a third loss function for the centripetal vector map, the third loss function being used to measure the difference between the predicted centripetal vector map and the real centripetal vector map, the third loss function being, for example, an L1 loss; performing weighted sum processing on the first loss function, the second loss function and the third loss function to obtain a total loss; and performing back propagation optimization according to the total loss to obtain the fourth feature map.

[0092] S5: generating a center heat map, an offset map and a centripetal vector map based on the target feature map through a prediction module.

[0093] In an optional embodiment, the target feature map is mapped to a single-channel heat map through a 1x1 convolutional layer, and the probability value of each pixel being the target center is output to obtain the center heat map; the fusion feature map is mapped to 2K channels (K is the number of key points) through a 1x1 convolutional layer to generate the x-axis and y-axis offset of each key point relative to the center point to obtain the offset map; and the fusion feature map is mapped to 2 channels through a 1x1 convolutional layer to generate the centripetal vector of each pixel, and the vector is normalized to obtain the centripetal vector map.

[0094] S6: extracting the center point coordinates of the vertebral body according to the center heat map, the offset map and the centripetal vector map, and determining the four corner point coordinates of the vertebral body according to the center point coordinates of the vertebral body.

[0095] Specifically, a non-maximum suppression algorithm is applied to the center heat map, and a preset number of peak points with the highest confidence score are selected in the processed center heat map. For each selected peak point, its integer coordinates on the target feature map and the confidence score are recorded. According to the integer coordinates of the peak point, a corresponding offset vector is extracted from the offset map, and the center point coordinates of the vertebra are determined according to the integer coordinates and the offset vector. According to the center point coordinates of the vertebra, eight vector components corresponding to the center point coordinates are extracted from the centripetal vector map, and the four corner point coordinates of the vertebra are determined according to the center point coordinates of the vertebra and the extracted vector components.

[0096] In an optional embodiment, the center heat map H heat The non-maximum suppression (NMS) algorithm is applied to eliminate redundant responses and retain local peak points. In the heat map processed by NMS, the K (for example, K = 17, corresponding to 12 thoracic vertebrae + 5 lumbar vertebrae, a total of 17) peak points p peak For each selected peak point i (i = 1,...,K), its integer coordinates on the feature map and the confidence score s i .

[0097] Further, the integer coordinates obtained in H heat are compensated by using the offset map O to obtain the center point coordinates with sub-pixel accuracy. Specifically, according to the integer coordinates of the peak point , a corresponding offset vector (Δx i ,Δy i ) is extracted from the offset map O. The accurate center point coordinates (x c,i ,y c,i ) are calculated according to the following formula:

[0098]

[0099] Further, the centripetal vector map V and the refined center point coordinates (x c,i ,y c,i ) are used to calculate the coordinates of the four corner points (for example, left upper TL, right upper TR, left lower BL, and right lower BR) of each vertebra. According to the integer coordinates of the peak point or the center point coordinates, eight vector components corresponding to the center point coordinates are extracted from the centripetal vector map V. The extracted components represent vectors from the four corner points to the center point, denoted as:

[0100] (v tl,x,i ,v tl,y,i ),(v tr,x,i ,v tr,y,i ),(v bl,x,i ,v bl,y,i ),(vbr,x,i ,v br,y,i )。

[0101] Since the predicted vector points from the corner point to the center point, the coordinates of each corner point are obtained by subtracting the corresponding vector components from the center point coordinates. For example, the calculation process of the coordinates (x tl,i ,y tl,i ) of the top-left corner point TL satisfies the following formula:

[0102] x tl,i =x c,i -v tl,x,i

[0103] x tl,i =x c,i -v tl,x,i

[0104] Similarly, the coordinates of the top-right TR, bottom-left BL, and bottom-right BR corner points can be calculated as follows:

[0105] x tr,i =x c,i -v tr,x,i ,y tr,i =y c,i -v tr,y,i

[0106] x bl,i =x c,i -v bl,x,i ,y bl,i =y c,i -v bl,y,i

[0107] x br,i =x c,i -v br,x,i ,y br,i =y c,i -v br,y,i

[0108] For each detected vertebral center point i, the four calculated corner point coordinates (x tl,i ,y tl,i ),(x tr,i ,y tr,i ),(x bl,i ,y bl,i ),(x br,i ,y br,i ) are combined as the final detection result of the key point, which is used for subsequent Cobb angle geometry calculation.

[0109] S7: Determine the Cobb angle by geometric algorithm according to the four corner point coordinates of the vertebra.

[0110] Specifically, the direction line of each vertebra is determined based on the coordinates of its four corner points; the angles between all pairs of vertebrae are determined based on the direction lines of the vertebrae, resulting in an angle matrix; the range of vertebrae corresponding to the main curve is determined, and the Cobb angle of the main curve is determined based on the angle matrix; the range of vertebrae corresponding to the secondary curve is determined, and the Cobb angle of the secondary curve is determined through a geometric algorithm.

[0111] In an optional embodiment, for each detected vertebral body k (k = 1, ..., N) v , where N v (This refers to the number of vertebrae), using the coordinates of their four corner points (x... tl,k ,y tl,k ),(x tr,k ,y tr,k ),(x bl,k ,y bl,k ),(x br,k ,y br,k The midpoints of its left and right edges are calculated according to the following formula:

[0112]

[0113] Where, p left_mid,k p represents the midpoint of the left edge. right_mid,k p represents the midpoint of the right edge. tl,k Represents the corner coordinates (x) tl,k ,y tl,k The coordinate vector of p) bl,k Represents the corner coordinates (x) bl,k ,y bl,k The coordinate vector of p) tr,k Represents the corner coordinates (x) tr,k ,y tr,k The coordinate vector of p) br,k Represents the corner coordinates (x) br,k ,y br,k The coordinate vector of ). The line segment connecting the two midpoints defines the direction line (or horizontal line of the central axis) of the vertebra k, and its direction vector The calculation process satisfies the following formula:

[0114]

[0115] Furthermore, calculate any two cones i and j (i, j = 1, ..., N) v ) direction vector and The included angle θ ij This angle represents the Cobb angle if vertebrae i and j are endplate vertebrae, and is calculated using the following formula:

[0116]

[0117] Where · represents the vector dot product, and ||·|| represents the Euclidean norm of the vectors. Calculate the angles between all pairs of cones to obtain an angle matrix Θ = [θ]. ij ].

[0118] Furthermore, in all calculated angles θ ij In the mean curve, find the angle with the largest absolute value; this angle is the Cobb angle θ of the principal curve (principal thoracic vertebra / thoracolumbar segment). MT :

[0119] θ MT =max i,j {θ ij}

[0120] Determine the two vertebral indices i that reach this maximum value. * and j * These correspond to the upper and lower endplate vertebrae of the main curve, respectively.

[0121] Furthermore, by analyzing the overall arrangement shape of all vertebral body center points (e.g., by assessing their deviation from the straight line connecting the center points of the first and last vertebrae), it can be determined whether the spine exhibits a distinct hyperbola ("S" shape) or is primarily a monobola ("C" shape).

[0122] For the case of a single curve: the Cobb angle θ of the proximal thoracic vertebrae PT It is usually calculated as the direction line of the first (uppermost) vertebra and the upper endplate vertebra i of the main curve. * The angle θ between the direction lines 1,i* Cobb angle θ of the thoracolumbar / lumbar spine TL It is usually calculated as the N of the last (lowest) vertebra. v Direction line and lower end plate vertebral body of the main curve j * Angle between direction lines .

[0123] For the hyperbola case: Typically, the terminal lamina i is located at the upper end of the principal curve. * Locate the endplate vertebral body of the proximal thoracic curve within the vertebral body above, through this region (e.g., vertebral bodies 1 to 1). * Find the maximum angle θ again. ij Determining the Cobb angle θ of the proximal thoracic vertebrae PT And its corresponding endplate vertebral body. Similarly, the endplate vertebral body of the thoracolumbar / lumbar curve will be located at the lower end of the main curve. * Search within the lower vertebral body, through this region (e.g., vertebral body j). * To N v Find the maximum angle to determine the Cobb angle θ of the thoracolumbar / lumbar spine.TL and the corresponding endplate vertebral body.

[0124] Finally, according to the positions of the upper and lower endplates in the spine, the three calculated angles are determined in the order from top to bottom as the Cobb angle θ of the proximal thoracic vertebrae PT , the Cobb angle θ of the main curve (main thoracic vertebrae / thoracolumbar segment) MT , and the Cobb angle θ of the thoracolumbar segment / lumbar vertebrae TL After obtaining the accurate Cobb angle, a quantitative reference can be provided for doctors to assist in the diagnosis and treatment decision of cervical spondylosis.

[0125] The method for determining the Cobb angle of scoliosis based on Mamba provided in the embodiments of the present application obtains image information, and pre-processes the obtained image information to obtain an original image; a first feature map is obtained by performing multi-level local feature extraction on the original image through an encoder; a second feature map is obtained by capturing global features of the first feature map through a Mamba module; a target feature map is obtained by decoding the second feature map through a decoder; a center heat map, an offset map and a centripetal vector map are generated based on the target feature map through a prediction module; the center point coordinates of the vertebral body are extracted according to the center heat map, the offset map and the centripetal vector map, and the four corner point coordinates of the vertebral body are determined according to the center point coordinates of the vertebral body; the Cobb angle is determined through a geometric algorithm according to the four corner point coordinates of the vertebral body, thereby improving the accuracy of Cobb angle measurement.

[0126] Figure 2 The structure diagram of the device for determining the Cobb angle of scoliosis based on Mamba provided in the embodiments of the present application is shown in FIG. 1. Figure 2 As shown in FIG. 1, the device 200 for determining the Cobb angle of scoliosis based on Mamba provided in the embodiments of the present application comprises:

[0127] The acquisition module 201 is configured to acquire image information.

[0128] The first processing module 202 is configured to pre-process the acquired image information to obtain an original image.

[0129] The second processing module 203 is configured to perform multi-level local feature extraction on the original image through an encoder to obtain a first feature map, wherein the encoder comprises a plurality of residual modules.

[0130] The second processing module 203 is further configured to capture global features of the first feature map through a Mamba module to obtain a second feature map.

[0131] The second processing module 203 is further configured to decode the second feature map through a decoder to obtain a target feature map, wherein the decoder comprises a main decoding branch and an auxiliary decoding branch.

[0132] The second processing module 203 is further configured to generate a center heat map, an offset map and a centripetal vector map based on the target feature map by a prediction module.

[0133] The second processing module 203 is further configured to extract the center point coordinates of the vertebral body according to the center heat map, the offset map and the centripetal vector map, and determine the four corner point coordinates of the vertebral body according to the center point coordinates of the vertebral body.

[0134] The second processing module 203 is further configured to determine the Cobb angle of the scoliosis by a geometric algorithm according to the four corner point coordinates of the vertebral body.

[0135] The device for determining the Cobb angle of scoliosis based on Mamba provided in the embodiment can perform the method for determining the Cobb angle of scoliosis based on Mamba provided in the method embodiment, and has similar implementation principles and technical effects, which will not be described here.

[0136] Figure 3 A structural schematic diagram of the device for determining the Cobb angle of scoliosis based on Mamba provided in the embodiment is shown in FIG. 3. Figure 3 As shown in FIG. 3, the device for determining the Cobb angle of scoliosis based on Mamba provided in the embodiment includes a receiver 301, a transmitter 302, a processor 303 and a memory 304.

[0137] The receiver 301 is configured to receive instructions and data.

[0138] The transmitter 302 is configured to transmit instructions and data.

[0139] The memory 304 is configured to store computer execution instructions.

[0140] The processor 303 is configured to execute the computer execution instructions stored in the memory 304 to implement each step performed by the method for determining the Cobb angle of scoliosis based on Mamba in the above embodiment. For details, refer to the related description in the method embodiment for determining the Cobb angle of scoliosis based on Mamba.

[0141] Optionally, the memory 304 can be independent or integrated with the processor 303.

[0142] When the memory 304 is independently arranged, the electronic device further includes a bus for connecting the memory 304 and the processor 303.

[0143] The embodiment of the present application further provides a computer storage medium, and the computer storage medium stores computer execution instructions. When a processor executes the computer execution instructions, the method for determining a Mamba-based scoliosis Cobb angle is implemented.

[0144] The embodiment of the present application further provides a computer program product, which comprises a computer program. When the computer program is executed by a processor, the method for determining a Mamba-based scoliosis Cobb angle is implemented.

[0145] Those skilled in the art can understand that all or some steps in the method disclosed above and the functional modules / units in the system and device can be implemented as software, firmware, hardware or appropriate combination thereof. In the hardware implementation, the division between the functional modules / units mentioned in the above description does not necessarily correspond to the division of physical components; for example, one physical component can have multiple functions, or one function or step can be performed by several physical components in cooperation. Some or all of the physical components can be implemented as software executed by a processor such as a central processing unit, a digital signal processor or a microprocessor, or as hardware, or as an integrated circuit such as an application specific integrated circuit. Such software can be distributed on a computer readable medium, which can include computer storage media (or non-transitory media) and communication media (or transitory media). As known to those skilled in the art, the term computer storage media includes volatile and non-volatile, removable and non-removable media implemented in any method or technology for storage of information such as computer readable instructions, data structures, program modules or other data. Computer storage media includes, but is not limited to, RAM, ROM, EEPROM, flash memory or other memory technology, CD-ROM, digital versatile disks (DVD) or other optical disk storage, magnetic cassettes, magnetic tapes, magnetic disk storage or other magnetic storage devices, or any other medium that can be used to store desired information and can be accessed by a computer. In addition, as known to those skilled in the art, communication media typically includes computer readable instructions, data structures, program modules or other data in modulated data signals such as carrier waves or other transport mechanisms, and can include any information delivery medium.

[0146] Other embodiments of the present application will be apparent to those skilled in the art from consideration of the specification and practice of the application disclosed herein. The specification and examples given are exemplary only. It is intended to cover any variations, uses or adaptations of the application following the general principles thereof contained within the scope of the application. The specification and examples are illustrative only and not restrictive of the true scope and spirit of the application, which is set forth in the following claims.

[0147] It is to be understood that the application is not limited to the precise construction herein described and as shown in the drawings, and that various modifications and changes can be made by those skilled in the art without departing from the scope thereof. The scope of the application is limited only by the appended claims.

Claims

1. A method for determining a Mamba-based scoliosis Cobb angle, the method comprising: The method comprises: acquiring image information and pre-processing the acquired image information to obtain an original image; extracting multi-level local features of the original image through an encoder to obtain a first feature map, wherein the encoder comprises a plurality of residual modules; capturing global features of the first feature map through a Mamba module to obtain a second feature map; the capturing of the global features of the first feature map through the Mamba module to obtain the second feature map comprises: performing feature serialization processing on the first feature map to obtain a first feature vector sequence; analyzing and processing the first feature vector sequence through a structured state space model of the Mamba module to obtain a second feature vector sequence; and mapping the second feature vector sequence back to a two-dimensional feature map to obtain the second feature map; decoding the second feature map through a decoder to obtain a target feature map, wherein the decoder comprises a main decoding branch and an auxiliary decoding branch; generating a center heat map, an offset map and a centripetal vector map based on the target feature map through a prediction module; extracting a center point coordinate of a vertebral body according to the center heat map, the offset map and the centripetal vector map, and determining four corner point coordinates of the vertebral body according to the center point coordinate of the vertebral body; determining a Cobb angle through a geometric algorithm according to the four corner point coordinates of the vertebral body.

2. The method of claim 1, wherein, The decoding of the second feature map through the decoder to obtain the target feature map comprises: performing up-sampling and convolution operations on the second feature map through the main decoding branch to obtain a third feature map; receiving features of an intermediate layer of the encoder through the auxiliary decoding branch, and performing back propagation optimization through an auxiliary loss function to obtain a fourth feature map; performing fusion processing on the third feature map and the fourth feature map to obtain the target feature map.

3. The method of claim 1, wherein, The extraction of the center point coordinate of the vertebral body according to the center heat map, the offset map and the centripetal vector map, and the determination of the four corner point coordinates of the vertebral body according to the center point coordinate of the vertebral body comprise: applying a non-maximum suppression algorithm to the center heat map for processing, and selecting a preset number of peak points with the highest confidence score in the processed center heat map, and recording the integer coordinates and the confidence score of each selected peak point on the target feature map; extracting corresponding offset vectors from the offset map according to the integer coordinates of the peak points, and determining the center point coordinate of the vertebral body according to the integer coordinates and the offset vectors; extracting corresponding eight vector components from the centripetal vector map according to the center point coordinate of the vertebral body, and determining the four corner point coordinates of the vertebral body according to the center point coordinate of the vertebral body and the extracted vector components.

4. The method of claim 1, wherein, The determination of the Cobb angle through the geometric algorithm according to the four corner point coordinates of the vertebral body comprises: determining a direction line of each vertebral body according to the four corner point coordinates of the vertebral body; determining angles between all pairs of vertebral bodies according to the direction lines of the vertebral bodies to obtain an angle matrix; determining a range of vertebral bodies corresponding to a main curve, and determining a Cobb angle of the main curve according to the angle matrix; determining a range of vertebral bodies corresponding to a secondary curve, and determining a Cobb angle of the secondary curve through a geometric algorithm.

5. The method of claim 2, wherein, The back propagation optimization through the auxiliary loss function to obtain the fourth feature map comprises: defining a first loss function for the center heat map, and the first loss function is used to measure the difference between the predicted center heat map and the real center heat map; a second loss function is defined for the offset map, the second loss function being used to measure a difference between the predicted offset map and the real offset map; a third loss function is defined for the centripetal vector map, the third loss function being used to measure a difference between the predicted centripetal vector map and the real centripetal vector map; the first loss function, the second loss function and the third loss function are processed by weighted summation to obtain a total loss; the fourth feature map is obtained by back propagation optimization according to the total loss.

6. A Mamba-based scoliosis Cobb angle determination apparatus, characterized in that, The device comprises: an acquisition module configured to acquire image information; a first processing module configured to pre-process the acquired image information to obtain an original image; a second processing module configured to extract multi-level local features of the original image by an encoder to obtain a first feature map, wherein the encoder comprises a plurality of residual modules; the second processing module is further configured to capture global features of the first feature map by a Mamba module to obtain a second feature map, and the capturing of the global features of the first feature map by the Mamba module to obtain the second feature map comprises: performing feature serialization processing on the first feature map to obtain a first feature vector sequence; analyzing and processing the first feature vector sequence by a structured state space model of the Mamba module to obtain a second feature vector sequence; and mapping the second feature vector sequence back to a two-dimensional feature map to obtain the second feature map; the second processing module is further configured to decode the second feature map by a decoder to obtain a target feature map, wherein the decoder comprises a main decoding branch and an auxiliary decoding branch; the second processing module is further configured to generate a center heat map, an offset map and a centripetal vector map based on the target feature map by a prediction module; the second processing module is further configured to extract a center point coordinate of a vertebral body according to the center heat map, the offset map and the centripetal vector map, and determine four corner point coordinates of the vertebral body according to the center point coordinate of the vertebral body; the second processing module is further configured to determine a Cobb angle of the vertebral body by a geometric algorithm according to the four corner point coordinates of the vertebral body.

7. A Mamba-based scoliosis Cobb angle determination device, characterized in that, The device comprises: a memory; a processor; wherein the memory stores computer execution instructions; the processor executes the computer execution instructions stored in the memory to implement the Mamba-based spinal scoliosis Cobb angle determination method according to any one of claims 1-5.

8. A computer storage medium, characterized in that The computer storage medium stores computer execution instructions, and the computer execution instructions are executed by the processor to implement the Mamba-based spinal scoliosis Cobb angle determination method according to any one of claims 1-5.

9. A computer program product, characterised in that, The computer program is executed by the processor to implement the Mamba-based spinal scoliosis Cobb angle determination method according to any one of claims 1-5.

Images