3D Medical Image Blurred Boundary Segmentation Method and System Based on Multi-Feature Perception

Abstract

The present invention relates to the field of 3D medical image segmentation technology, and specifically relates to a 3D medical image fuzzy boundary segmentation method and system based on multi-feature perception, as follows: preprocess the 3D medical image to be segmented to obtain a low-dimensional feature sequence; construct a medical 3D shape capture attention module to process the feature sequence to obtain shape attention features; construct a 3D image boundary perception segmentation network, input the feature sequence into the 3D image boundary perception segmentation network for feature extraction to obtain boundary perception features; construct a 3D image boundary perception discriminator, splice and fuse the shape attention features and the boundary perception features to obtain a segmentation result, then input it into the 3D image boundary perception discriminator for evaluation, and then optimize the 3D image boundary perception segmentation network according to the evaluation result feedback. The present invention can capture the fine heterogeneous texture features of 3D medical images through a multi-feature perception attention module, make full use of the multi-level information of 3D medical images, and improve the segmentation effect.

Description

Technical Field

[0001] The present invention relates to the technical field of 3D medical image segmentation, and in particular to a 3D medical image fuzzy boundary segmentation method and system based on multi-feature perception. Background Art

[0002] With the development of medical information technology, 3D imaging technology has played a significant role in clinical practice. 3D medical imaging diagnostics are primarily used for early diagnosis, lesion localization, and treatment planning. They provide detailed three-dimensional structural information of organs, allowing examiners to accurately observe the morphology, size, and location of organs, tumors, or other lesions, thereby helping doctors develop personalized treatment plans. 3D images can also monitor disease progression, helping doctors improve the accuracy of clinical decision-making.

[0003] However, 3D medical images may involve three-dimensional spatial information of multiple different tissues or organs. The examiner needs to analyze, extract, and process the details of each slice. Therefore, only by using 3D image-assisted segmentation technology can the efficiency of inspection and analysis be effectively improved. With the development of artificial intelligence technology, deep learning has gradually been applied to the field of 3D medical image segmentation. Since the boundaries of 3D medical images may be fuzzy or complex in shape, general 3D medical image segmentation methods based on deep learning cannot effectively and accurately segment the image boundaries. Traditional threshold methods and region growing methods rely on image grayscale differences, while fuzzy areas often show characteristics of gentle gradients and overlapping intensities, resulting in broken or over-segmented segmentation results. In addition, active contour models are easily affected by the initial position and easily fall into local optimality in complex anatomical structures. The core challenges of fuzzy boundaries in medical images are heterogeneity, deformation, and low signal-to-noise ratio.

[0004] Therefore, the present invention proposes a 3D medical image fuzzy boundary segmentation method and system based on multi-feature perception to solve the above problems. Summary of the Invention

[0005] In response to the shortcomings of the existing technology, the present invention develops a 3D medical image fuzzy boundary segmentation method and system based on multi-feature perception. The present invention can capture the subtle heterogeneous texture features of 3D medical images through a multi-feature perception attention module, make full use of the multi-level information of 3D medical images, and improve the segmentation effect.

[0006] On the one hand, the technical solution to the technical problem of the present invention is a 3D medical image fuzzy boundary segmentation method based on multi-feature perception, comprising the following steps:

[0007] S1. Preprocess the 3D medical image to be segmented to obtain a low-dimensional feature sequence;

[0008] S2. Construct a medical 3D morphological capture attention module, input the low-dimensional feature sequence into the medical 3D morphological capture attention module to obtain morphological attention features;

[0009] S3. Construct a 3D image boundary-aware segmentation network, which includes two multi-feature perception attention modules MFPA and a boundary-aware module BAM. Input the low-dimensional feature sequence into the 3D image boundary-aware segmentation network for feature extraction to obtain boundary-aware features;

[0010] S4. Construct a 3D image boundary-aware discriminator, splice and fuse the morphological attention features and the boundary-aware features to obtain a segmentation result, then input the segmentation result into the 3D image boundary-aware discriminator for evaluation, and then optimize the 3D image boundary-aware segmentation network according to the evaluation result feedback.

[0011] S1 is specifically as follows:

[0012] Obtain the 3D medical image to be segmented from the existing dataset. Any 3D medical image to be segmented is represented as , , represents the height, represents the width, represents the depth, represents the number of channels;

[0013] Construct a 3D medical image preprocessing module, and input the 3D medical image to be segmented into the 3D medical image preprocessing module. Among them, through a sliding window of size sliding in the image , the image is divided into feature blocks, and each feature block is represented as , , , the calculation process of dividing the image into feature blocks is as follows:

[0014] ,

[0015] Among them, represents the height of the sliding window, represents the width of the sliding window, represents the depth of the sliding window;

[0016] Then, flatten the feature block into a one-dimensional vector , and the calculation formula is as follows:

[0017] ,

[0018] Among them, represents the th feature block flattened into a corresponding one-dimensional vector , indicating the flattening operation;

[0019] By linearly transforming the one-dimensional vector and projecting it into a low-dimensional feature space, a feature vector with a fixed length is obtained. , represents the dimension after projection, and the calculation formula is as follows:

[0020] ,

[0021] Among them, represents the linear projection matrix, , represents the bias vector, ;

[0022] Finally, feature vectors with a fixed length are combined into a feature sequence , , .

[0023] S2 is specifically as follows:

[0024] Construct a medical 3D shape capture attention module, which includes a normalization layer, a variable shape capture attention module, a 3D convolutional layer with a convolutional kernel size and a 3D convolutional layer with a convolutional kernel size of . Input the low-dimensional feature sequence output by the 3D medical image preprocessing module into the medical 3D shape capture attention module;

[0025] The feature sequence is first normalized by the normalization layer to obtain the normalized feature , and then the normalized feature is input into the variable shape capture attention module. The normalized feature passes through a 3D convolutional layer with a convolutional kernel size of and[[ID=7vi]] and activation function to obtain the feature , and the calculation formula is as follows:

[0026] ;

[0027] Then, the feature The convolution kernel size in the attention module is captured in a variable form. 3D depth-wise separable convolutional layer , the convolution kernel size is 3D dilated convolutional layer And the convolution kernel size is 3D convolutional layer , get the convolution feature , and then the convolution feature and features Splice and output features , and then the features The convolution kernel size input to the variable shape capture attention module is The 3D convolution layer combines the obtained features with the normalized features Splicing is performed to obtain the output features of the variable form capture attention module , the calculation formula is as follows:

[0028] ,

[0029] ,

[0030] ,

[0031] in, Represents a splicing operation;

[0032] Next, the output features of the attention module are captured by the variable form With low-dimensional feature sequences Features are spliced to obtain enhanced features after information supplementation , strengthen the characteristics After convolution kernel size 3D convolutional layer And the convolution kernel size is 3D convolutional layer , get the features ,feature and enhanced features Perform feature splicing to finally obtain the output morphological attention features of the medical 3D morphological capture attention module , the calculation formula is as follows:

[0033] ,

[0034] ,

[0035] .

[0036] S3 is as follows:

[0037] The two multi-feature perception attention modules MFPA are the first multi-feature perception attention module and the second multi-feature perception attention module respectively. Both of the two multi-feature perception attention modules include a spatial attention feature capture module and a channel attention capture module inside. The input features pass through the spatial attention feature capture module and the channel attention capture module respectively to obtain spatial attention features and channel attention features, and then the two features are fused to obtain the output of the feature perception attention module;

[0038] The boundary perception module splices and fuses the outputs of the two feature perception attention modules, forms a boundary perception key feature map according to the intersection over union of the boundary region and the 3D true segmentation boundary image, and then calculates the boundary perception features from the spliced and fused features and the boundary perception key feature map.

[0039] The specific operations of the multi-feature perception attention module are as follows:

[0040] Input the feature sequence into the first multi-feature perception attention module to obtain the first multi-feature perception feature , and then input the first multi-feature perception feature into the second multi-feature perception attention module to obtain the second multi-feature perception feature . The first multi-feature perception attention module and the second multi-feature perception attention module share the weights of the query and the key;

[0041] The specific operation process in the first multi-feature perception attention module is as follows:

[0042] (1) The morphological attention feature passes through the spatial attention feature capture module:

[0043] Calculate the shared query , the shared key and the spatial value , and the calculation formulas are as follows:

[0044] ,

[0045] ,

[0046] ,

[0047] where , , represent the projection matrices of the query, the key, and the value respectively;

[0048] Then, through the shared query and the shared key Calculate the dot product to compute the similarity between features, obtaining the spatial attention feature map , then multiply the spatial attention feature map by the spatial value to obtain the spatial attention feature . The calculation formula is as follows:

[0049] ,

[0050] ,

[0051] where, represents the activation function, represents the shared key in its transposed form, represents the dimension of the shared query and key;

[0052] (2) The feature sequence passes through the channel attention feature capture module:

[0053] ,

[0054] ,

[0055] ,

[0056] where, and represent the query and key of the channel attention module, represents the channel value;

[0057] Compute the similarity between features through the dot product of the shared query and the shared key to obtain the channel attention feature map , then multiply the channel attention feature map by the channel value to obtain the channel attention feature . The calculation formula is as follows:

[0058] ,

[0059] ;

[0060] (3) Add and fuse the spatial attention feature and the channel attention feature to obtain the fused feature . Input the fused feature sequentially into the convolutional layer with a convolutional kernel size of and the convolutional kernel size of Convolution layer In it, the final first multi-feature perception attention module feature is obtained , and the calculation formula is as follows:

[0061] ,

[0062] ;

[0063] The specific operation in the second multi-feature perception attention module is the same as that in the first multi-feature perception attention module. The first multi-feature perception attention module feature is input into the second multi-feature perception attention module to obtain the second multi-feature perception attention module feature .

[0064] The operation in the boundary perception module is specifically as follows:

[0065] The first multi-feature perception attention module feature and the second multi-feature perception attention module feature are concatenated and fused to generate the perception fusion feature ;

[0066] Using the Canny edge detection algorithm, the boundary of the target object is extracted from the corresponding real 3D medical segmentation image of the 3D medical image to be segmented . The target boundary contour of the 3D image is obtained, and randomly selected on the target boundary contour of the 3D image candidate boundary key points to obtain the key point set , represents the th iteration of randomly selected boundary point sets, . According to the key point set , the corresponding boundary region is constructed, represents the th key point, , , and respectively represent the x-axis, y-axis, and z-axis coordinates of the [[ID=6)9]]th randomly selected key point on the target boundary contour of the 3D image in the th iteration;

[0067] Calculate the intersection over union of the boundary region and the real 3D medical segmentation image , and the calculation formula is as follows:

[0068] ,

[0069] Among them, represents the set of optimal boundary points obtained after calculating the intersection over union (IoU), which contains candidate boundary key points. According to the set of optimal boundary points construct the corresponding boundary-aware key feature map ;

[0070] Fuse the key feature map with the perception fusion feature using channel weighting operation to obtain the boundary-aware feature , and the calculation process is as follows:

[0071] ,

[0072] Among them, represents element-wise addition, represents channel-wise multiplication.

[0073] The process of calculating the segmentation result is specifically as follows:

[0074] Fuse the boundary-aware feature with the morphological attention feature in a concatenated manner to fuse them in the channel dimension to obtain the fused feature . Adjust the number of channels of the image using a convolutional layer with a convolutional kernel size of and use activation function to output the final predicted segmentation 3D image , and the calculation formula is as follows:

[0075] ,

[0076] ,

[0077] Among them, represents the concatenation operation of features, represents the convolutional operation with a convolutional kernel size of .

[0078] The 3D image boundary-aware discriminator is as follows:

[0079] If the structure boundary of the target region retained in the segmentation 3D image finally predicted by the 3D image boundary-aware segmentation network is the same as the real 3D medical segmentation image ​If they are consistent, the 3D image boundary perception discriminator outputs a high score; conversely, if the true segmentation image fails to match, a low score is output; among them, the discriminator module is to evaluate the predicted segmentation result output by the evaluation model To evaluate the quality of the discriminator loss function, design the discriminator loss function so that the discriminator module can distinguish between high-score segmentation images and mis-segmented images. The calculation formula is as follows:

[0080] ,

[0081] ,

[0082] Among them, represents the discriminator function, represents the score output by the discriminator, represents the concatenation operation, represents a convolutional layer with a convolutional kernel size of ; represents the fully connected layer, represents the Sigmoid activation function, represents the discriminator loss function, represents the logarithmic function with base 2;

[0083] Optimize the 3D image boundary perception segmentation network with the score output by the discriminator, introduce the 3D image boundary loss, and adjust its own parameters according to the score. The calculation formula is as follows:

[0084] ,

[0085] Among them, represents the 3D image boundary loss, represents the logarithmic function with base 2.

[0086] On the other hand, the present invention also provides a 3D medical image fuzzy boundary segmentation system based on multi-feature perception, including a module for executing the processing instructions of each step in the 3D medical image fuzzy boundary segmentation method based on multi-feature perception;

[0087] Image preprocessing module: Preprocess the 3D medical image to be segmented to obtain a low-dimensional feature sequence;

[0088] Medical 3D morphology capture attention module: Input the low-dimensional feature sequence into the medical 3D morphology capture attention module to obtain morphological attention features;

[0089] 3D image boundary perception segmentation module: Includes two multi-feature perception attention modules MFPA and one boundary perception module BAM. Input the morphological attention features into the 3D image boundary perception segmentation module for feature extraction to obtain boundary perception features;

[0090] 3D Image Boundary Perception Discriminant Module: Concatenate and fuse the morphological attention features and boundary perception features to obtain the segmentation result, then input the segmentation result into the 3D image boundary perception discriminator for evaluation, and then feedback and optimize the 3D image boundary perception segmentation network according to the evaluation result.

[0091] The effects provided in the invention content are only the effects of the embodiments, rather than all the effects of the invention. The above technical solutions have the following advantages or beneficial effects:

[0092] The present invention discloses a 3D medical image fuzzy boundary segmentation method and system based on multi-feature perception. The multi-feature perception attention module is used to capture the subtle heterogeneous texture features of 3D medical images. Combining with the boundary perception module, the captured features are used to enhance the model's exploration of the fuzzy boundaries of objects, making full use of the multi-level information of 3D medical images, namely spatial-channel dual-path attention, and sharing weights, which can take into account both global and local features. It also actively learns high-confidence boundary key points through intersection over union screening and channel weighted fusion; the constructed medical 3D morphological capture attention module uses the deformable convolution mechanism to dynamically calibrate the modal feature heteroscedastic noise through the normalization layer and degree-scale feature concatenation, and effectively models the long-range dependence relationship and local structure deformation information between features in three-dimensional space. Through the combination of depthwise separable convolution and dilated convolution, the computational complexity is significantly reduced, and the model's expressiveness for complex organ boundaries and irregular morphologies is enhanced; at the same time, the present invention also introduces a 3D medical image boundary perception discriminator to specifically optimize the fuzzy region, score the 3D image after the model segmentation, and feedback the scoring result to the model, effectively improving the segmentation effect of the 3D medical image fuzzy boundary. BRIEF DESCRIPTION OF THE DRAWINGS

[0093] The drawings are used to provide a further understanding of the present invention, and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention, and do not constitute a limitation to the present invention.

[0094] Figure 1 It is a schematic flowchart of the method of the present invention.

[0095] Figure 2 It is a structural diagram of the system of the present invention.

[0096] Figure 3 It is a comparison chart of the segmentation effects of the method of the present invention and the existing method. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0097] To clearly illustrate the technical features of this solution, the present invention will be elaborated in detail below through specific embodiments and in conjunction with its accompanying drawings. The following disclosure provides many different embodiments or examples for implementing different structures of the present invention. To simplify the disclosure of the present invention, components and settings of specific examples are described below.

[0098] Embodiment 1

[0099] A 3D medical image fuzzy boundary segmentation method based on multi-feature perception, comprising the following steps:

[0100] S1. Preprocess the 3D medical image to be segmented to obtain a low-dimensional feature sequence;

[0101] S2. Construct a medical 3D morphological capture attention module, input the low-dimensional feature sequence into the medical 3D morphological capture attention module to obtain morphological attention features;

[0102] S3. Construct a 3D image boundary perception segmentation network, which includes two multi-feature perception attention modules MFPA and a boundary perception module BAM, input the low-dimensional feature sequence into the 3D image boundary perception segmentation network for feature extraction to obtain boundary perception features;

[0103] S4. Construct a 3D image boundary perception discriminator, splice and fuse the morphological attention features and the boundary perception features to obtain a segmentation result, then input the segmentation result into the 3D image boundary perception discriminator for evaluation, and then optimize the 3D image boundary perception segmentation network according to the evaluation result feedback.

[0104] S1 is specifically as follows:

[0105] Obtain the 3D medical image to be segmented from the existing dataset. Any 3D medical image to be segmented is represented as , , represents the height, represents the width, represents the depth, represents the number of channels;

[0106] Construct a 3D medical image preprocessing module, input the 3D medical image to be segmented into the 3D medical image preprocessing module. Among them, through a sliding window of size sliding in the image , the image is divided into feature blocks, and each feature block is represented as , , , the image is divided into The calculation process of a feature block is as follows:

[0107] ,

[0108] Among them, represents the height of the sliding window, represents the width of the sliding window, represents the depth of the sliding window;

[0109] Then, the feature block is flattened into a one-dimensional vector , and the calculation formula is as follows:

[0110] ,

[0111] Among them, represents the th feature block after being flattened into the corresponding one-dimensional vector , represents the flattening operation;

[0112] The one-dimensional vector is projected into a low-dimensional feature space through a linear transformation to obtain a feature vector with a fixed length, , represents the dimension after projection, and the calculation formula is as follows:

[0113] ,

[0114] Among them, represents the linear projection matrix, , represents the bias vector, ;

[0115] Finally, fixed-length feature vectors are combined into a feature sequence , , .

[0116] S2 is specifically as follows:

[0117] Construct a medical 3D shape capture attention module, which includes a normalization layer, a variable shape capture attention module, a 3D convolutional layer with a convolutional kernel size of and a 3D convolutional layer with a convolutional kernel size of , and input the low-dimensional feature sequence output by the 3D medical image preprocessing module into the medical 3D shape capture attention module;

[0118] Feature Sequence First, normalization is performed through the normalization layer to obtain normalized features , and then input the normalized features into the variable form capture attention module, the normalized features After the convolution kernel size is 3D convolutional layer and Activation function gets features , the calculation formula is as follows:

[0119] ;

[0120] Then, the features The convolution kernel size in the attention module is captured in a variable form. 3D depthwise separable convolutional layer , the convolution kernel size is 3D dilated convolutional layer And the convolution kernel size is 3D convolutional layer , get the convolution feature , and then the convolution feature and features Splice and output features , and then the features The convolution kernel size input to the variable shape capture attention module is The 3D convolution layer combines the obtained features with the normalized features Splicing is performed to obtain the output features of the variable form capture attention module , the calculation formula is as follows:

[0121] ,

[0122] ,

[0123] ,

[0124] in, Represents a splicing operation;

[0125] Next, the output features of the attention module are captured by the variable form With low-dimensional feature sequences Features are spliced to obtain enhanced features after information supplementation , strengthen the characteristics After convolution kernel size 3D convolutional layer And the convolution kernel size is 3D convolutional layer , obtain the feature , the feature and the enhanced feature are concatenated by features, and finally the morphological attention feature output by the medical 3D morphological capture attention module is obtained. The calculation formula is as follows:

[0126] ,

[0127] ,

[0128] .

[0129] S3 is specifically as follows:

[0130] The two multi-feature perception attention modules MFPA are the first multi-feature perception attention module and the second multi-feature perception attention module respectively. Both of the two multi-feature perception attention modules include a spatial attention feature capture module and a channel attention capture module inside. The input features pass through the spatial attention feature capture module and the channel attention capture module respectively to obtain the spatial attention feature and the channel attention feature, and then the two features are fused to obtain the output of the feature perception attention module;

[0131] The boundary perception module splices and fuses the outputs of the two feature perception attention modules, forms a boundary perception key feature map according to the intersection over union of the boundary region and the 3D true segmentation boundary image, and then calculates the boundary perception feature from the spliced and fused features and the boundary perception key feature map.

[0132] The specific operation of the multi-feature perception attention module is as follows:

[0133] Input the feature sequence into the first multi-feature perception attention module to obtain the first multi-feature perception feature , and then input the first multi-feature perception feature into the second multi-feature perception attention module to obtain the second multi-feature perception feature The first multi-feature perception attention module and the second multi-feature perception attention module share the weights of the query and the key;

[0134] The specific operation process in the first multi-feature perception attention module is as follows:

[0135] (1) The morphological attention feature passes through the spatial attention feature capture module:

[0136] Calculate the shared query , the shared key and the spatial value , and the calculation formula is as follows:

[0137] ,

[0138] ,

[0139] ,

[0140] in, 、 、 The projection matrices representing query, key, and value respectively;

[0141] Then, by sharing the query and Share Key The dot product calculation is used to calculate the similarity between features and obtain the spatial attention feature map , and then the spatial attention feature map With space value Multiply to get the spatial attention feature , the calculation formula is as follows:

[0142] ,

[0143] ,

[0144] in, represents the activation function, Indicates a shared key The transpose of Dimensions representing shared queries and keys;

[0145] (2) Feature sequence After the channel attention feature capture module:

[0146] ,

[0147] ,

[0148] ,

[0149] in, and represents the query and key of the channel attention module, Indicates channel value;

[0150] By sharing the query and Share Key The dot product calculation is used to calculate the similarity between features and obtain the channel attention feature map , and then the channel attention feature map With channel value Multiply to get channel attention features , the calculation formula is as follows:

[0151] ,

[0152] ;

[0153] (3) Add and fuse the spatial attention feature and the channel attention feature to obtain the fused feature . Input the fused feature successively into the convolutional layer with a convolutional kernel size of and the convolutional layer with a convolutional kernel size of and to obtain the final feature of the first multi - feature perception attention module . The calculation formula is as follows:

[0154]

[0155] ,

[0156] ; The specific operations in the second multi - feature perception attention module are the same as those in the first multi - feature perception attention module. Input the feature of the first multi - feature perception attention module into the second multi - feature perception attention module to obtain the feature of the second multi - feature perception attention module.

[0157] The operations in the boundary perception module are as follows:

[0158] Concatenate and fuse the feature of the first multi - feature perception attention module and the feature of the second multi - feature perception attention module to generate the perceptual fusion feature ;

[0159] Use the Canny edge detection algorithm to extract the boundaries of the target objects from the corresponding real 3D medical segmentation image of the 3D medical image to be segmented . Obtain the target boundary contour of the 3D image. Randomly select candidate boundary key points on the target boundary contour of the 3D image to obtain the key point set . represents the - th iteration of randomly selected boundary point sets. . According to the key point set , construct the corresponding boundary region . Indicates the th key point, , , and respectively represent the x-axis, y-axis, and z-axis coordinates of the th key point randomly selected on the target boundary contour of the 3D image in the th iteration;

[0160] Calculate the intersection over union of the boundary region with the true 3D medical segmentation image , and the calculation formula is as follows:

[0161] ,

[0162] where, represents the set of optimal boundary points obtained after calculating the intersection over union, containing candidate boundary key points, and construct the corresponding boundary-aware key feature map according to the set of optimal boundary points ;

[0163] Fuse the key feature map with the perception fusion feature using channel weighting operation to obtain the boundary-aware feature , and the calculation process is as follows:

[0164] ,

[0165] where, represents element-wise addition, represents channel-wise multiplication.

[0166] The process of calculating the segmentation result is as follows:

[0167] Fuse the boundary-aware feature with the morphological attention feature in a concatenated manner and fuse them in the channel dimension to obtain the fused feature , adjust the number of channels of the image using a convolutional layer with a convolutional kernel size of and use activation function to output the final predicted segmentation 3D image , and the calculation formula is as follows:

[0168]

[0169] ,

[0170] ​​​Among them, represents the splicing operation of features, represents that the convolution kernel size is for the convolution operation.

[0171] The 3D image boundary-aware discriminator is as follows:

[0172] If the structure boundary of the target region retained in the segmented 3D image finally predicted by the 3D image boundary-aware segmentation network is consistent with the real 3D medical segmentation image , the 3D image boundary-aware discriminator outputs a high score; otherwise, if it fails to match the real segmentation image, it outputs a low score; among them, the discriminator module is to evaluate the quality of the predicted segmentation result output by the model, and a discriminator loss function is designed so that the discriminator module can distinguish between high-score segmentation images and wrongly segmented images. The calculation formula is as follows:

[0173] ,

[0174] ,

[0175] Among them, represents the discriminator function, represents the score output by the discriminator, represents the splicing operation, represents that the convolution kernel size is for the convolutional layer, represents the fully connected layer, represents the Sigmoid activation function, represents the discriminator loss function, represents the logarithmic function with base 2;

[0176] The 3D image boundary-aware segmentation network is optimized with the score output by the discriminator, and the 3D image boundary loss is introduced to adjust its own parameters according to the score. The calculation formula is as follows:

[0177] ,

[0178] Among them, represents the 3D image boundary loss, represents the logarithmic function with base 2.

[0179] Example 2

[0180] A 3D medical image fuzzy boundary segmentation system based on multi-feature perception includes a module for executing the processing instructions of each step in the 3D medical image fuzzy boundary segmentation method based on multi-feature perception; <000>

[0181] ​Image preprocessing module: Preprocess the 3D medical image to be segmented to obtain a low-dimensional feature sequence;

[0182] Medical 3D shape capture attention module: Input the low-dimensional feature sequence into the medical 3D shape capture attention module to obtain shape attention features;

[0183] 3D image boundary-aware segmentation module: It includes two multi-feature perception attention modules MFPA and one boundary awareness module BAM. Input the shape attention features into the 3D image boundary-aware segmentation module for feature extraction to obtain boundary awareness features;

[0184] 3D image boundary-aware discrimination module: Concatenate and fuse the shape attention features and the boundary awareness features to obtain a segmentation result, then input the segmentation result into the 3D image boundary-aware discriminator for evaluation, and then optimize the 3D image boundary-aware segmentation network according to the evaluation result feedback.

[0185] Embodiment 3

[0186] In the verification experiment of the present invention, four experimental metrics are used to verify the effectiveness of the proposed method, namely HD95mm, DSC, Params, and FLOPs; HD95mm represents the boundary difference between the predicted segmentation image result and the ground truth label. The lower the value, the closer the boundary predicted by the model is to the true boundary; DSC (Dice similarity coefficient) is used to measure the overlap degree between the predicted segmentation result and the manually segmented label. The higher the DSC value, the higher the overlap degree between the segmentation result of the model and the ground truth label; Params represents the number of parameters of the model, reflecting the complexity of the model. Under the same segmentation result, the model with fewer parameters is better; FLOPs represents the number of floating-point operations, used to measure the computational complexity of the model. The lower the FLOPs value, the higher the computational efficiency of the model.

[0187] Under the same experimental conditions and environmental configurations, the proposed 3D medical image fuzzy boundary segmentation method based on multi-feature perception was compared with other 3D image segmentation methods to verify the effectiveness of the proposed method. The other comparison models are as follows: The classic U-Net image segmentation model extracts and fuses multi-scale features through convolution and deconvolution operations. It is suitable for basic image segmentation tasks, but the receptive field of the model structure is limited, and its ability to segment images in complex scenarios is weak. The Swin UNet is a 3D image segmentation model based on the Swin Transformer. It realizes the fusion of local and global features of 3D images through the window attention mechanism, but the computational complexity of the model is relatively high. The nnFormer is an improved hybrid image segmentation model that uses an attention module to implement a feature pyramid structure. The performance of this algorithm is relatively excellent, but its computational complexity and the number of parameters are relatively high. The UNETR uses a Transformer as the encoder to capture the long-range dependencies of 3D image features, and the decoder is based on a 3D convolutional network. This model is suitable for 3D image segmentation tasks, but the complexity of the model is relatively high.

[0188] The proposed 3D medical image fuzzy boundary segmentation method based on multi-feature perception was verified on the dataset. The experimental results are shown in Table 1. In the performance comparison with other 3D image segmentation models, the comprehensive performance of the proposed method in this invention is more excellent. Among them, the HD95mm index reached the lowest value of 10.57, indicating that the error between the predicted 3D object image boundary and the real boundary is the smallest. The DSC index of the proposed method is 83.64, which is significantly higher than that of other segmentation models, indicating that the segmentation accuracy of the model is the best. In the two indicators of Params and FLOPs, the scores of the model are 61.49 and 75.80 respectively. The scores of these two indicators show that the model has better performance with a lower number of parameters and computational cost.

[0189] The experimental results show that the method proposed in this invention has advantages in both accuracy and efficiency, verifying its effectiveness in 3D image segmentation tasks.

[0190] Table 1 Comparison results between the method of this invention and existing methods

[0191]

[0192] To more intuitively demonstrate that this invention has a better 3D medical image fuzzy boundary segmentation effect, the method of this invention was compared with two existing methods, Swin Unet and U-Net, as Figure 3As shown in the figure, it is a comparison chart of the segmentation effects of three groups of the method of the present invention and the existing method. By comparing the segmentation effects of the method of the present invention, Swin-Unet and U-Net with the real segmentation image, it can be directly seen that the coincidence degree between the segmentation result area obtained by the method of the present invention and the real segmentation area is higher. Thus, it can be further proved that the method in the present invention is effective in the 3D image segmentation task.

[0193] Although the specific implementation manners of the invention are described above in conjunction with the accompanying drawings, it is not a limitation on the protection scope of the present invention. Based on the technical solutions of the present invention, various modifications or deformations that can be made by those skilled in the art without creative efforts are still within the protection scope of the present invention.

Claims

1. A 3D medical image fuzzy boundary segmentation method based on multi-feature perception, characterized in that It includes the following steps: S1. Preprocess the 3D medical image to be segmented to obtain a low-dimensional feature sequence; S2. Construct a medical 3D morphological capture attention module, which includes a normalization layer, a variable morphological capture attention module, a 3D convolutional layer with a convolutional kernel size and a 3D convolutional layer with a convolutional kernel size of . Input the low-dimensional feature sequence into the medical 3D morphological capture attention module to obtain morphological attention features; S3. Construct a 3D image boundary-aware segmentation network, which includes two multi-feature perception attention modules MFPA and a boundary-aware module BAM. Input the low-dimensional feature sequence into the 3D image boundary-aware segmentation network for feature extraction to obtain boundary-aware features; The two multi-feature perception attention modules MFPA are the first multi-feature perception attention module and the second multi-feature perception attention module respectively. Both multi-feature perception attention modules internally include a spatial attention feature capture module and a channel attention capture module. The input features pass through the spatial attention feature capture module and the channel attention capture module respectively to obtain spatial attention features and channel attention features, and then the two features are fused to obtain the output of the feature perception attention module; S4. Construct a 3D image boundary-aware discriminator. Concatenate and fuse the morphological attention features and the boundary-aware features to obtain a segmentation result, then input the segmentation result into the 3D image boundary-aware discriminator for evaluation, and then optimize the 3D image boundary-aware segmentation network according to the evaluation result feedback.

2. The 3D medical image fuzzy boundary segmentation method based on multi-feature perception according to claim 1, characterized in that, S1 is specifically as follows: Obtain the 3D medical image to be segmented from the existing dataset, and any 3D medical image to be segmented is represented as , , represents the height, represents the width, represents the depth, represents the number of channels; Build a 3D medical image preprocessing module for the 3D medical image to be segmented Input it into the 3D medical image preprocessing module, where A sliding window of size Slides in the image The image Is divided into Feature blocks, each feature block is represented as , , the image Is divided into The calculation process of the feature blocks is as follows: ， Among them, represents the height of the sliding window, represents the width of the sliding window, represents the depth of the sliding window; Then, the feature block is flattened through a flattening operation into a one-dimensional vector , and the calculation formula is as follows: ， Among them, represents the th feature block flattened into the corresponding one-dimensional vector , indicating the flattening operation; When projecting a one-dimensional vector through a linear transformation into a low-dimensional feature space, a feature vector of a fixed length is obtained , , represents the dimension after projection, and the calculation formula is as follows: ， Among them, represents a linear projection matrix, , represents a bias vector, ; Finally, fixed-length feature vectors are combined into a feature sequence , , .

3. A 3D medical image fuzzy boundary segmentation method based on multi-feature perception according to claim 2, characterized in that S2 Specifically as follows: Construct a medical 3D morphological capture attention module and input the low-dimensional feature sequence output by the 3D medical image preprocessing module into the medical 3D morphological capture attention module; Feature sequence First, it is normalized by a normalization layer to obtain normalized features , and then the normalized features are input into a variable form capture attention module. The normalized features pass through a 3D convolutional layer with a convolutional kernel size of and an activation function to obtain features . The calculation formula is as follows: ​ ； Then, the features The convolution kernel size in the attention module is captured in a variable form. 3D depthwise separable convolutional layer , the convolution kernel size is 3D dilated convolutional layer And the convolution kernel size is 3D convolutional layer , get the convolution feature , and then the convolution feature and features Splice and output features , and then the features The convolution kernel size input to the variable shape capture attention module is The 3D convolution layer combines the obtained features with the normalized features Splicing is performed to obtain the output features of the variable form capture attention module , the calculation formula is as follows: ， ， ， Among them, represents a splicing operation; Next, the output features of the variable form capture attention module are concatenated with the low-dimensional feature sequence features to obtain the enhanced features after information supplementation . The enhanced features pass through a 3D convolutional layer with a convolutional kernel size of and a 3D convolutional layer with a convolutional kernel size of to obtain the features . The features are concatenated with the enhanced features to finally obtain the morphological attention features output by the medical 3D form capture attention module. The calculation formula is as follows:​​ ， ， 。 4. A 3D medical image fuzzy boundary segmentation method based on multi-feature perception according to claim 3, characterized in that S3 Specifically as follows: The boundary-aware module concatenates and fuses the outputs of the two feature perception attention modules, constructs a boundary-aware key feature map according to the intersection over union of the boundary region and the 3D true segmentation boundary image, and then calculates the boundary-aware features from the concatenated and fused features and the boundary-aware key feature map.

5. A 3D medical image fuzzy boundary segmentation method based on multi-feature perception according to claim 4, characterized in that The specific operations of the multi-feature perception attention module are as follows: Input the feature sequence into the first multi-feature perception attention module to obtain the first multi-feature perception feature . Then input the first multi-feature perception feature into the second multi-feature perception attention module to obtain the second multi-feature perception feature . The first multi-feature perception attention module and the second multi-feature perception attention module share the weights of the query and the key; The specific operation process in the first multi-feature perception attention module is as follows: (1)Morphological attention feature After the spatial attention feature capture module: Calculated shared query and shared key as well as spatial value , and the calculation formula is as follows: ， ， ， Among them, , , respectively represent the projection matrices for query, key, and value; Then, through the dot product calculation of the shared query and the shared key to calculate the similarity between features, obtaining the spatial attention feature map , and then multiplying the spatial attention feature map by the spatial value to obtain the spatial attention feature . The calculation formula is as follows: ， ， Among them, represents the activation function, represents the transpose of the shared key and represents the dimension of the shared query and key; (2) Feature sequence Pass through the channel attention feature capture module: ， ， ， Among them, and represent the query and key of the channel attention module, represents the channel value; Through shared queries and shared keys calculate the similarity between features through the dot product calculation to obtain the channel attention feature map , and then multiply the channel attention feature map by the channel value to obtain the channel attention feature . The calculation formula is as follows: ， ； (3) Spatial attention features and channel attention features Add and fuse to obtain the fused features , the fused features Input into the convolution kernel size in sequence The convolutional layer And the convolution kernel size is The convolutional layer In the final first multi-feature perception attention module feature , the calculation formula is as follows: ， ； The specific operations in the second multi-feature perception attention module are the same as those in the first multi-feature perception attention module. The features of the first multi-feature perception attention module are input into the second multi-feature perception attention module to obtain the features of the second multi-feature perception attention module .

6. A 3D medical image fuzzy boundary segmentation method based on multi-feature perception according to claim 5, characterized in that, The operations in the boundary-aware module are specifically as follows: Concatenate and fuse the features of the first multi-feature perception attention module and the features of the second multi-feature perception attention module to generate the perception fusion features ; Using the Canny edge detection algorithm on the 3D medical image to be segmented The corresponding real 3D medical segmentation image Extract the boundaries of the target object, obtain the target boundary contour of the 3D image, and randomly select candidate boundary key points on the target boundary contour of the 3D image to obtain the key point set , indicating the th time of randomly selected boundary point sets, , construct the corresponding boundary region according to the key point set , , indicating the th key point, , , and respectively represent the x-axis, y-axis and z-axis coordinates of the th key point randomly selected on the target boundary contour of the 3D image in the th iteration; Calculating the boundary region with the true 3D medical segmentation image Intersection over Union , and the calculation formula is as follows: ， Among them, represents the set of optimal boundary points obtained after calculating the intersection over union (IoU), which contains candidate boundary key points. According to the set of optimal boundary points the corresponding boundary-aware key feature map is constructed ; Fuse the key feature map with the perception fusion feature using a channel weighting operation to obtain a boundary perception feature , and the calculation process is as follows: ， Among them, represents element-wise addition, represents channel-wise multiplication.

7. A 3D medical image fuzzy boundary segmentation method based on multi-feature perception according to claim 6, characterized in that, The process of calculating the segmentation result is specifically as follows: The boundary-aware features and the morphological attention features are fused in the channel dimension by concatenation to obtain the fused features . The fused features are adjusted in the number of image channels using a convolutional layer with a convolutional kernel size of , and the final predicted segmented 3D image is output using the activation function. The calculation formula is as follows: ​ ， ， Among them, represents the splicing operation of features, represents that the convolution kernel size is convolution operation of 8. A 3D medical image fuzzy boundary segmentation method based on multi-feature perception according to claim 7, characterized in that, The 3D image boundary-aware discriminator is specifically as follows: If the segmented 3D image finally predicted by the 3D image boundary-aware segmentation network has consistent structural boundaries of the retained target regions with the true 3D medical segmentation image , the 3D image boundary-aware discriminator outputs a high score; conversely, if it fails to match the true segmentation image, it outputs a low score; among them, the discriminator module is to evaluate the quality of the predicted segmentation result output by the model . To design a discriminator loss function to enable the discriminator module to distinguish between high-score segmentation images and mis-segmented images, the calculation formula is as follows: ， ， Among them, represents the discriminator function, represents the score output by the discriminator, represents the concatenation operation, represents that the convolutional kernel size is of the convolutional layer, represents the fully connected layer, represents the Sigmoid activation function, represents the discriminator loss function, represents the logarithm function with base 2; Optimize the 3D image boundary-aware segmentation network with the score output by the discriminator, introduce the 3D image boundary loss to adjust its own parameters according to the score, and the calculation formula is as follows: ， Among them, represents the 3D image boundary loss, represents the logarithm function with base 2.

9. A 3D medical image fuzzy boundary segmentation system based on multi-feature perception, which executes a 3D medical image fuzzy boundary segmentation method according to any one of claims 1-8, characterized in that, It includes the following modules: Image preprocessing module: Preprocess the 3D medical image to be segmented to obtain a low-dimensional feature sequence; Medical 3D morphology capture attention module: Input the low-dimensional feature sequence into the medical 3D morphology capture attention module to obtain morphological attention features; 3D image boundary-aware segmentation module: It includes two multi-feature perception attention modules MFPA and a boundary-aware module BAM. Input the morphological attention features into the 3D image boundary-aware segmentation module for feature extraction to obtain boundary-aware features; 3D image boundary-aware discrimination module: Concatenate and fuse the morphological attention features and the boundary-aware features to obtain a segmentation result, then input the segmentation result into the 3D image boundary-aware discriminator for evaluation, and then optimize the 3D image boundary-aware segmentation network according to the evaluation result feedback.

Images