Pigtail wishbone three-dimensional irregular rejection interface generation method

Abstract

The invention provides a method for generating a three-dimensional irregular rejecting interface of a pigtail wishbone. The method comprises the following steps: acquiring a three-dimensional curvature model of the pigtail wishbone; based on the three-dimensional curvature model, according to the complexity of the shape of the tail wishbone, carrying out partitioning processing on the tail wishbone region to obtain regions with similar complexity, and recording the regions as blocks; a self-adaptive three-dimensional segmentation model is constructed for different areas with similar complexity, and an optimal tail wishbone removal interface is obtained; the mode that the tail wishbone is divided into blocks and then fused is adopted, not only can the varied forms of bones be flexibly coped, the personalized deboning effect be achieved, but also the continuity and smoothness of a deboning interface can be ensured, subsequent automatic deboning operation is facilitated, and meat loss is reduced. In addition, the adaptive three-dimensional segmentation model performs specific processing according to the complexity of different bone blocks, so that the segmentation efficiency is optimized, and the segmentation accuracy is improved.

Description

Technical Field

[0001] The invention relates to the field of livestock carcass segmentation, and in particular to a method for generating a three-dimensional irregular elimination interface of a pigtail wishbone. Background Art

[0002] Pork is rich in nutrients and is widely used in various dishes and food processing. It is an indispensable part of global food culture. In the pork production process, fine segmentation is a key step to ensure meat quality and economic benefits. Among them, the tail fork bone has a complex geometric structure, variable and irregular shape. It has always been a complex and challenging link to remove it from the pig's hind leg. At present, the removal of the tail fork bone in China still relies on manual work. In the process of manual deboning, even if there is a certain amount of segmentation experience, due to the complexity and invisibility of the bone structure, there will still be problems such as damage to the key soft tissues closely connected to the tail fork bone, tool sliding and jamming, and difficulty in maintaining the shape requirements of the tail fork bone after deboning, which leads to low deboning quality and efficiency. Although the use of mechanical automation to debone can reduce the labor burden of workers and improve the efficiency of deboning, it is difficult to obtain the precise deboning interface between the pig's tail fork bone and the surrounding tissue before the machine performs the deboning operation, resulting in the inability of the tool to segment along the predetermined trajectory, resulting in meat loss.

[0003] The invention patent with application number 202111393371.7 discloses a deboning and segmentation device and method for livestock carcasses. The deboning and segmentation device includes: a workbench, a bone scanning mechanism, a segmentation and deboning robot, and an auxiliary grasping robot. The deboning and segmentation method is: fix the livestock carcass to be segmented on the workbench; move the workbench to the scanning position, and use the X-ray in the bone scanning mechanism to obtain the bone shape and position information; move the workbench to the cutting position, and start the grasping device to assist in fixing the bone; the segmentation and deboning robot automatically adjusts the cutting path according to the scanned bone information to complete the segmentation; and puts the separated bones and meat into the material basket respectively. Specifically, the shape, density, position and two-dimensional image data of the pig bones are captured by the X-ray scanning device in the bone scanning mechanism, and the data are converted into the three-dimensional shape information of the bones through an algorithm to form a bone model. The current bone model is compared with the standard bone model to obtain the deviation value between the actual bone and the standard bone, and the deviation value is used to correct the preset deboning path and motion trajectory of the segmentation and deboning robot. The method is simple in process, easy to control and implement, and improves accuracy and stability to a certain extent. However, over-reliance on standard models reduces flexibility in handling non-standard shapes. In the event of large deviations or sudden changes in the bone structure, the segmentation path may be offset, resulting in meat loss. In addition, the method is based on the deviation calculation of the overall model, rather than fine-tuning and optimizing the segmentation path for complex local areas in the bone structure, which affects the consistency of the culling effect. Summary of the invention

[0004] In view of the problems in the prior art, the present invention provides a method for generating a three-dimensional irregular removal interface of a pig's tail furcula, the purpose of which is to process the morphology of complex areas of the tail furcula and obtain an accurate tail furcula removal interface.

[0005] A method for generating a three-dimensional irregular elimination interface of a pigtail wishbone comprises the following steps: Step 1: Obtain a three-dimensional curvature model of the coccygeal wishbone; Step 2: Based on the three-dimensional curvature model, the coccyx region is divided into blocks according to the complexity of the shape of the coccyx, and regions with similar complexity are obtained and recorded as blocks; Step 3: For regions of similar complexity, an adaptive 3D segmentation model is constructed to obtain the optimal forkbone removal interface.

[0006] Further, the adaptive three-dimensional segmentation model includes a first network unit, a second network unit and an overall adaptive three-dimensional segmentation network, wherein the first network unit is used to process the complexity score data of each block of the coccyx region to form a complexity score processing layer, which is then introduced into a key position of the second network unit, and a dynamic feature modeling path is generated through a Mamba module according to the distribution of the complexity score data and the input feature map; The second network unit is used to process the three-dimensional point cloud data of the coccyx region, and then gradually extract and restore features to perform preliminary segmentation of the coccyx region; The overall adaptive three-dimensional segmentation network is composed of a first network unit and a second network unit. The second network unit combines the complex scoring processing layer and the Mamba module in the first network unit to segment and refine the features of each block in the forkbone area to generate the final accurate forkbone removal interface.

[0007] Further, the first network unit is composed of two parts: the complexity score processing layer and the Mamba module; The complexity score processing layer processes the complexity score data through a fully connected layer (FC) and expands the score into a tensor form that matches the dimension of the feature map of the key position of the second network unit; The Mamba module consists of a linear transformation (Linear), a 1D convolution (Conv1D), a nonlinear activation function (SiLU), and a selective state space (S6). In the Mamba module, the input feature map 𝑥 first undergoes two parallel linear transformations to generate two different feature streams. Secondly, the first feature stream is subjected to a 1D convolution operation to capture the local coccygeal features and extract the local feature patterns and short-term dependencies of the features. Then, both feature streams pass through the Silu activation function. The formula of the SiLU activation function is:

[0008] in, It is the Sigmoid function, which enhances the sensitivity to the input value; Then, the first feature stream continues to pass through the selective state space; finally, the two feature streams are added through the residual connection to obtain the fused features, and the fused features are then passed through the linear layer to generate the final output features.

[0009] Further, the Mamba module dynamically adjusts the feature graph generated by a specific path through the selective state space, that is, the generated dynamic feature modeling path.

[0010] Further: the second network unit is composed of a 3D U-Net framework; The 3D U-Net framework consists of 3 encoding layers, 1 bottleneck layer and 3 decoding layers. Each encoding layer consists of two 3D convolutional layers (Conv). Each 3D convolutional layer is followed by batch normalization (BN) and Relu activation function (Conv+BN+Relu). At the end of each encoding layer, a 3D maximum pooling (Max pool) operation is used to halve the resolution of the feature map. During the encoding process of the encoding layer, the number of feature channels of the second network unit is doubled. The bottleneck layer contains two consecutive 3D convolution operations; Each decoding layer contains a deconvolution operation and is used to upsample the feature map. The feature map formed after upsampling is concatenated (Cat) with the corresponding encoding layer feature map through a jump connection to combine the low-level and high-level feature maps; the concatenated feature map undergoes two 3D convolution layers, batch normalization and Relu activation function operations; a separate 3D convolution operation is added to the last decoding layer.

[0011] Further, there are three key positions of the second network unit, namely: the position after the Max pool in the third encoding layer, the position after the Up Conv in the first decoding layer, and the position after the Up Conv in the third decoding layer.

[0012] Further: the overall adaptive 3D segmentation network consists of 3 encoding layers, 1 bottleneck layer and 3 decoding layers; The first two coding layers of the overall adaptive three-dimensional segmentation network are the same as the first two coding layers in the second network unit. The third coding layer of the overall adaptive three-dimensional segmentation network includes the same structure as the third coding layer of the second network unit, and in the same structure as the third coding layer of the second network unit, the score in the complexity score processing layer is expanded to a tensor that is the same as the feature map generated after the Max pool operation. Then, the complexity score processing layer and the feature map generated after the Max pool operation are broadcasted, and the feature map generated after the broadcast operation is the deep feature map of the third coding layer. The broadcast operation introduces the key position after the Max pool in the third coding layer of the second network unit into the complexity score processing layer of this part. Finally, the deep feature map generated after the broadcast operation of the third coding layer is further selectively modeled through the Mamba module, and the feature expression ability of the network is enhanced through local convolution and state space mechanisms, so as to adapt to the feature requirements of different complexity areas, and form the third coding layer of the overall adaptive three-dimensional segmentation network. The bottleneck layer of the adaptive 3D segmentation network is used to connect the encoding layer and the decoding layer, and is the same as the bottleneck layer in the second network unit; The first decoding layer of the overall adaptive three-dimensional segmentation network includes the same structure as the first decoding layer of the second network unit. In the same structure as the first decoding layer of the second network unit, the score in the complexity score processing layer will be expanded to the same tensor as the feature map generated after the deconvolution operation; then, the complexity score processing layer and the feature map generated after the deconvolution operation are broadcasted, and the feature map generated after the broadcast operation is the deep feature map of the first decoding layer of the adaptive three-dimensional segmentation network. The broadcast operation introduces the complexity score processing layer of this part after the deconvolution in the first decoding layer of the second network unit; then, the deep feature map generated after the broadcast operation of the first decoding layer is further selectively modeled through the Mamba module, and the feature expression ability of the network is enhanced through local convolution and state space mechanisms, so as to adapt to the feature requirements of different complexity areas; finally, after two 3D convolution layers, batch normalization and Relu activation function operations (Conv+BN+Relu 2) After that, the first decoding layer of the adaptive 3D segmentation network is formed; The second decoding layer of the adaptive 3D segmentation network is the same as the second decoding layer in the second network unit; The third decoding layer of the overall adaptive three-dimensional segmentation network includes the same structure as the third decoding layer of the second network unit. In the same structure as the first decoding layer of the second network unit, the score in the complexity score processing layer will be expanded to the same tensor as the feature map generated after the deconvolution operation; then the complexity score processing layer and the feature map generated after the deconvolution operation are broadcasted, and the feature map generated after the broadcast operation is the deep feature map of the first decoding layer of the adaptive three-dimensional segmentation network. The broadcast operation introduces the complexity score processing layer of this part after the deconvolution in the first decoding layer of the second network unit; then, the Mamba module further selectively models the deep feature map generated after the broadcast operation of the first decoding layer, and enhances the feature expression ability of the network through local convolution and state space mechanisms, so as to adapt to the feature requirements of different complexity areas; finally, after two 3D convolution layers, batch normalization and Relu activation function operations (Conv+BN+Relu 2) After that, the third decoding layer of the adaptive 3D segmentation network is formed; Furthermore, the adaptive 3D segmentation model is trained using a cross entropy loss function, an additional regularization term balancing model, and a gradient descent optimization algorithm.

[0013] Further, step 1 comprises the following steps: Step 1.1: Based on the slice-by-slice images of the cross-section of the three-dimensional model of the pig hind leg, a threshold segmentation technique is used to obtain a preliminary segmentation line of the pig hind leg caudal wishbone; Step 1.2: Perform curvature analysis on the preliminary segmentation line to quantify the curvature of the bone in each slice image and obtain the curvature value of the preliminary segmentation line on each slice image; Step 1.3: Use the multi-directional slice curvature value data mapping method to construct a three-dimensional curvature model of the coccygeal wishbone.

[0014] Further: Step 2.1: Map the curvature data in each slice image in each direction to the three-dimensional space, obtain the three-dimensional point cloud containing curvature information in each direction, and set the number of block categories , the 3D point cloud is divided into blocks; For a 3D point cloud, the feature vector of each point is composed of its 3D spatial position and curvature value The eigenvector is ; First, sort the curvature values ​​of all points. After sorting, the threshold can be determined directly based on the statistical distribution of the curvature values; after sorting the curvature values ​​from small to large, select multiple appropriate quantiles as the division The threshold of the complexity level area is set. According to the set threshold, the position of each point in the three-dimensional space of the coccyx is matched with its curvature category to form Point sets of different complexity, composed of Different block areas, i.e. blocks; Step 2.2: Use K-Means++ method to initialize cluster centers; Randomly select a data point from the data set formed by the 3D point cloud as the first cluster center; calculate the Euclidean distance between each data point in the data set and the cluster center; for each data point that is not selected as a cluster center, calculate its distance to the nearest cluster center; assign the probability of being selected as the next cluster center based on the square of these distances, the larger the distance, the higher the probability of being selected; use the probability distribution calculated in the previous step to select a new cluster center by random sampling; repeat the above steps until the selection is completed. The selected N cluster centers are the initial center points; Each time the K-Means++ algorithm selects a new center, it calculates the distance and probability based on the currently selected center and all data points. Suppose the coordinates of the data point P in the three-dimensional curvature model are , the coordinates of cluster center C are ; Step 2.3: Assign categories to the three-dimensional curvature model to obtain preliminary classification results; Assign each data point to the "cluster" where the nearest initial center point is located, forming Classification: Each part of the coccyx is classified into the group with the most similar complexity for preliminary classification; the category assignment process is as follows: for each data point , calculate its relationship with each cluster center Distance ,Will Assigned to the cluster center closest to it; Step 2.4: Update the cluster center, that is, according to the preliminary classification results, recalculate the center point of each cluster, that is, calculate the average position of all points in each category as the new cluster center; Step 2.5: Iterative optimization, that is, repeating steps 2.3 and 2.4, when the position of the cluster center tends to be stable or reaches the predetermined number of iterations, obtain The result of clusters is formed A sub-block composed of points of the same curvature category, that is, an area with similar complexity.

[0015] The beneficial effects of the present invention are as follows: the method of fusing the coccyx after processing the blocks can not only flexibly cope with the changeable forms of the bones and achieve a personalized deboning effect, but also ensure the continuity and smoothness of the deboning interface, facilitate the subsequent automated deboning operation, and reduce meat loss. In addition, the adaptive three-dimensional segmentation model performs specific processing according to the complexity of different bone blocks, which not only optimizes the segmentation efficiency but also improves the segmentation accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 It is a schematic diagram of the process of the present invention; Figure 2 This is the network structure diagram of the adaptive 3D segmentation model. DETAILED DESCRIPTION

[0017] The present invention is described in detail below in conjunction with the accompanying drawings. The embodiments of the present invention are described in detail below, and examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements with the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and cannot be interpreted as limiting the present invention. The directional terms such as left, middle, right, top, and bottom in the embodiments of the present invention are only relative concepts or are based on the normal use state of the product, and should not be considered as restrictive.

[0018] A method for generating a three-dimensional irregular elimination interface of a pigtail wishbone, such as Figure 1 As shown, the following steps are included: Step 1: Obtain a three-dimensional curvature model of the coccygeal wishbone; Step 1.1: Based on the slice-by-slice images of the cross-section of the three-dimensional model of the pig hind leg, a threshold segmentation technique is used to obtain a preliminary segmentation line of the pig hind leg caudal wishbone; Step 1.2: Perform curvature analysis on the preliminary segmentation line to quantify the curvature of the bone in each slice image and obtain the curvature value of the preliminary segmentation line on each slice image; Step 1.3: Using the multi-directional slice curvature value data mapping method, a three-dimensional curvature model of the coccygeal furcula is constructed; Step 2: Based on the three-dimensional curvature model, the coccyx region is divided into blocks according to the complexity of the shape of the coccyx, and regions with similar complexity are obtained and recorded as blocks; Step 2.1: Map the curvature data in each slice image in each direction to the three-dimensional space, obtain the three-dimensional point cloud containing curvature information in each direction, and set the number of block categories. , the 3D point cloud is divided into blocks; Before clustering, the data of the 3D bone model needs to be selected and prepared for features; for the 3D point cloud, the feature vector of each point is composed of its 3D spatial position and curvature value The eigenvector is ; First, sort the curvature values ​​of all points. After sorting, the threshold can be determined directly based on the statistical distribution of the curvature values; first, sort the curvature values ​​from small to large, and then select multiple appropriate quantiles, such as the 10th, 30th, and 50th quantiles as the division For example, when the number of categories is 2, the 50th percentile (i.e., median) is selected as the threshold to divide the curvature point set into two parts: high complexity area and low complexity area. When the number of categories is 4, the 25th, 50th, and 75th percentiles are selected as thresholds to divide the curvature point set into four parts with different complexities. According to the set threshold, the position of each point in the three-dimensional space of the coccyx is matched with its curvature category to form a Point sets of different complexity, composed of Different block areas, i.e. blocks; Step 2.2: Use K-Means++ method to initialize cluster centers and reduce the impact of outliers; In the three-dimensional curvature model of the coccygeal fork, the appearance of curvature outliers may be caused by pathological changes, trauma or other atypical features of the bones; therefore, the present invention uses K-Means++ to select the initial cluster center to reduce the impact of outliers. The steps of the K-Means++ algorithm are as follows: randomly select a data point from the data set formed by the three-dimensional point cloud as the first cluster center; calculate the Euclidean distance between each data point in the data set and the cluster center; for each data point that is not selected as a cluster center, calculate its distance to the nearest cluster center; assign the probability of being selected as the next cluster center based on the square of these distances, the greater the distance, the higher the probability of being selected; use the probability distribution calculated in the previous step to select a new cluster center by random sampling; repeat the above steps until a new cluster center is selected. The selected N cluster centers are the initial center points; Each time the K-Means++ algorithm selects a new center, it calculates the distance and probability based on the currently selected center and all data points. This method significantly improves the stability and clustering quality of the algorithm. Suppose the coordinates of the data point P in the three-dimensional curvature model are , the coordinates of cluster center C are ; Step 2.3: Assign categories to the three-dimensional curvature model to obtain preliminary classification results; Assign each data point to the "cluster" where the nearest initial center point is located, forming Classification: Each part of the coccyx is classified into the group with the most similar complexity for preliminary classification; the category assignment process is as follows: for each data point , calculate its relationship with each cluster center Distance ,Will Assigned to the cluster center closest to it; Step 2.4: Update the cluster center, that is, according to the preliminary classification results, recalculate the center point of each cluster, that is, calculate the average position of all points in each category as the new cluster center; Step 2.5: Iterative optimization, that is, repeating steps 2.3 and 2.4, when the position of the cluster center tends to be stable or reaches the predetermined number of iterations, obtain The result of clusters is formed sub-blocks consisting of points of the same curvature category, i.e., areas of similar complexity; Step 3: For regions of similar complexity, an adaptive 3D segmentation model is constructed to obtain the optimal forkbone removal interface; Among them, in order to solve the problem of complex and diverse bone structures and significant differences in regional characteristics during the process of forkbone removal, the existing segmentation methods usually adopt a unified calculation strategy, which is difficult to take into account the efficiency of simple areas and the refinement requirements of complex areas. The present invention proposes an adaptive three-dimensional segmentation model; this model is a model that can dynamically adjust the segmentation strategy according to the complexity of the region to achieve an efficient and refined segmentation model. The adaptability of this model is mainly the adaptation of the complexity of the forkbone model. The calculated complexity score of each block of the forkbone model After that, the model can adjust the segmentation strategy according to the complexity of each block. For areas with lower scores, the model uses a shallow network, fewer feature extraction layers or smaller convolution kernels to reduce the waste of computing resources. For areas with higher scores, the model uses a deeper network structure and more feature extraction modules to increase the model's ability to capture details. This adaptability is achieved by dynamically adjusting the feature extraction depth of the network, avoiding a "one-size-fits-all" calculation method for all areas, and can improve the segmentation efficiency of the model while ensuring segmentation accuracy. The adaptive 3D segmentation model is trained using the cross entropy loss function, an additional regularization term balancing model, and a gradient descent optimization algorithm. Combination Figure 2 As shown, the adaptive three-dimensional segmentation model includes a first network unit, a second network unit and an overall adaptive three-dimensional segmentation network; The first network unit is used to process the complexity score data of each block in the forkbone region to form a complexity score processing layer, which is then introduced into the key position of the second network unit, and a dynamic feature modeling path is generated through the Mamba module according to the distribution of the complexity score data and the input feature map; the first network unit consists of two parts: the complexity score processing layer and the Mamba module; the complexity score processing layer processes the complexity score data through a fully connected layer (FC) to expand the score into a tensor form that matches the dimension of the feature map at the key position of the second network unit; the Mamba module consists of a linear transformation (Linear), a convolution (Conv1D), a nonlinear activation function (SiLU) and a selective state space (S6); in the Mamba module, the input feature map 𝑥 first undergoes two parallel linear transformations to generate two different feature streams; secondly, the first feature stream captures the local forkbone features through a 1D convolution operation to extract the local feature patterns and short-term dependencies of the features; then, both feature streams pass through the Silu activation function to increase the nonlinear expression ability of the features. The formula of the SiLU activation function is:

[0019] in, It is the Sigmoid function, which enhances the sensitivity to the input value; Then, the first feature stream continues to pass through the selective state space, which controls the retention or update of the state through the gating mechanism, captures the global tail wishbone features, and enhances the dynamic modeling ability of the model; finally, the two feature streams are added through the residual connection to obtain the fused features, and the fused features are then passed through the linear layer to generate the final output features; The Mamba module dynamically adjusts the feature graph generated by a specific path through the selective state space, that is, the dynamic feature modeling path generated; The second network unit is used to process the three-dimensional point cloud data of the forkbone area, and then gradually extract and restore features to perform preliminary segmentation of the forkbone area. The overall adaptive three-dimensional segmentation network is composed of the first network unit and the second network unit. The second network unit combines the complex scoring processing layer and the Mamba module in the first network unit to segment and refine the features of each block in the forkbone area, and generate a final accurate forkbone removal interface. The second network unit is composed of a 3D U-Net framework. The 3D U-Net framework includes 3 encoding layers, 1 bottleneck layer and 3 decoding layers. Each encoding layer consists of two 3D convolutional layers (Conv). Each 3D convolutional layer is followed by batch normalization (BN) and Relu activation function (Conv+BN+Relu). At the end of each encoding layer, 3D maximum pooling (Max pool) operation is used to halve the resolution of the feature map. During the encoding process of the encoding layer, the number of feature channels of the second network unit is doubled; the bottleneck layer is used to perform high-level feature extraction and global modeling in the lowest resolution feature space to capture the global context information of the coccygeal region. The bottleneck layer contains two consecutive 3D convolution operations. The receptive field of the convolution kernel is large, which is used to capture more complex spatial features; the decoder is used to gradually upsample the low-resolution features extracted by the bottleneck layer, restore the high-resolution feature map, and generate the final segmentation result; each decoding layer contains a deconvolution (Up Conv) operation and used to upsample the feature map. The upsampled feature map is concatenated (Cat) with the corresponding encoding layer feature map through skip connections to combine the low-level and high-level feature maps; the concatenated feature map undergoes two 3D convolution layers, batch normalization, and Relu activation function operations; a separate 3D convolution operation is added to the last decoding layer; the restored high-resolution feature map is further processed to optimize the boundaries and details of the segmentation result to ensure that the output resolution is consistent with the input; Among them, there are three key positions of the second network unit, namely: the position after the Max pool in the third encoding layer, the position after the Up Conv in the first decoding layer, and the position after the Up Conv in the third decoding layer; The overall adaptive 3D segmentation network consists of 3 encoding layers, 1 bottleneck layer and 3 decoding layers; The overall adaptive three-dimensional segmentation network is based on the classic 3D U-Net structure, combining the embedding of the Mamba module and the complexity score 𝑆 to achieve accurate segmentation of the block data of the coccyx area; the overall network consists of an encoder, a bottleneck layer and a decoder, and through the collaborative work of the first network unit and the second network unit, it dynamically optimizes feature extraction and recovery, and finally generates a high-resolution segmentation result; the first two encoding layers of the overall adaptive three-dimensional segmentation network are the same as the first two encoding layers in the second network unit, and the third encoding layer of the overall adaptive three-dimensional segmentation network includes the same structure as the third encoding layer of the second network unit, and in the same structure as the third encoding layer of the second network unit, the score in the complexity score processing layer is expanded to the same tensor as the feature map generated after the Max pool operation at the position after the Max pool operation; then the complexity score processing layer and the feature map generated after the Max pool operation are broadcasted, and the feature map generated after the broadcast operation is the deep feature map of the third encoding layer, and the broadcast operation makes the Max in the third encoding layer of the second network unit The complexity scoring processing layer of this part is introduced at the key position after the pool; finally, the deep feature map generated after the broadcast operation of the third coding layer is further selectively modeled through the Mamba module, and the feature expression ability of the network is enhanced through local convolution and state space mechanism, so as to adapt to the feature requirements of different complexity areas and form the third coding layer of the overall adaptive 3D segmentation network; The bottleneck layer of the adaptive 3D segmentation network is used to connect the encoding layer and the decoding layer. It is mainly responsible for global feature modeling of the lowest resolution feature map and is the same as the bottleneck layer in the second network unit. The first decoding layer of the overall adaptive three-dimensional segmentation network includes the same structure as the first decoding layer of the second network unit. In the same structure as the first decoding layer of the second network unit, the score in the complexity score processing layer will be expanded to the same tensor as the feature map generated after the deconvolution operation; then the complexity score processing layer and the feature map generated after the deconvolution operation are broadcasted, and the feature map generated after the broadcasting operation is the deep feature map of the first decoding layer of the adaptive three-dimensional segmentation network. The broadcasting operation introduces the complexity score processing layer of this part after the deconvolution in the first decoding layer of the second network unit; then, the deep feature map generated after the broadcasting operation of the first decoding layer is further selectively modeled through the Mamba module, and the feature expression ability of the network is enhanced through local convolution and state space mechanisms, so as to adapt to the feature requirements of different complexity areas; finally, after two 3D convolution layers, batch normalization and Relu activation function operations (Conv+BN+Relu 2) After that, the first decoding layer of the adaptive 3D segmentation network is formed; The second decoding layer of the adaptive 3D segmentation network is the same as the second decoding layer in the second network unit; The third decoding layer of the overall adaptive three-dimensional segmentation network includes the same structure as the third decoding layer of the second network unit. In the same structure as the first decoding layer of the second network unit, the score in the complexity score processing layer will be expanded to the same tensor as the feature map generated after the deconvolution operation; then the complexity score processing layer and the feature map generated after the deconvolution operation are broadcasted, and the feature map generated after the broadcast operation is the deep feature map of the first decoding layer of the adaptive three-dimensional segmentation network. The broadcast operation introduces the complexity score processing layer of this part after the deconvolution in the first decoding layer of the second network unit; then, the Mamba module further selectively models the deep feature map generated after the broadcast operation of the first decoding layer, and enhances the feature expression ability of the network through local convolution and state space mechanisms, so as to adapt to the feature requirements of different complexity areas; finally, after two 3D convolution layers, batch normalization and Relu activation function operations (Conv+BN+Relu 2) After that, the third decoding layer of the adaptive 3D segmentation network is formed.

[0020] The above shows and describes the basic principles, main features and advantages of the present invention. It should be understood by those skilled in the art that the present invention is not limited to the above embodiments. The above embodiments and descriptions are only for explaining the principles of the present invention. Without departing from the spirit and scope of the present invention, the present invention may have various changes and improvements, which fall within the scope of the present invention. The scope of protection of the present invention is defined by the attached claims and their equivalents.

Claims

1. A method for generating a three-dimensional irregular elimination interface of a pigtail wishbone, characterized in that: The following steps are involved: Step 1: Obtain a three-dimensional curvature model of the coccygeal wishbone; Step 2: Based on the three-dimensional curvature model, the coccyx region is divided into blocks according to the complexity of the shape of the coccyx, and regions with similar complexity are obtained and recorded as blocks; Step 3: For regions of similar complexity, an adaptive 3D segmentation model is constructed to obtain the optimal forkbone removal interface.

2. The method for generating a three-dimensional irregular elimination interface of a pigtail wishbone according to claim 1, characterized in that: The adaptive 3D segmentation model includes a first network unit, a second network unit and an overall adaptive 3D segmentation network. The first network unit is used to process the complexity score data of each block in the caudal forkbone region to form a complexity score processing layer, which is then introduced into the key position of the second network unit, and a dynamic feature modeling path is generated through the Mamba module according to the distribution of the complexity score data and the input feature map. The second network unit is used to process the three-dimensional point cloud data of the coccyx region, and then gradually extract and restore features to perform preliminary segmentation of the coccyx region; The overall adaptive three-dimensional segmentation network is composed of a first network unit and a second network unit. The second network unit combines the complex scoring processing layer and the Mamba module in the first network unit to segment and refine the features of each block in the forkbone area to generate the final accurate forkbone removal interface.

3. The method for generating a three-dimensional irregular elimination interface of a pigtail wishbone according to claim 2, characterized in that: The first network unit consists of two parts: the complexity score processing layer and the Mamba module; The complexity score processing layer processes the complexity score data through a fully connected layer and expands the score into a tensor form that matches the dimension of the feature map of the key position of the second network unit; The Mamba module consists of linear transformation, 1D convolution, nonlinear activation function and selective state space. In the Mamba module, the input feature map 𝑥 first undergoes two parallel linear transformations to generate two different feature streams. Secondly, the first feature stream is subjected to a 1D convolution operation to capture the local coccygeal features and extract the local feature patterns and short-term dependencies of the features. Then, both feature streams pass through the Silu activation function. The formula of the SiLU activation function is: ； in, is the Sigmoid function; Then, the first feature stream continues to pass through the selective state space; finally, the two feature streams are added through the residual connection to obtain the fused features, and the fused features are then passed through the linear layer to generate the final output features.

4. The method for generating a three-dimensional irregular elimination interface of a pigtail wishbone according to claim 3, characterized in that: The Mamba module dynamically adjusts the feature maps generated by a specific path through the selective state space, that is, the generated dynamic feature modeling path.

5. The method for generating a three-dimensional irregular elimination interface of a pigtail wishbone according to claim 4, characterized in that: The second network unit is composed of a 3D U-Net framework; The 3D U-Net framework consists of 3 encoding layers, 1 bottleneck layer and 3 decoding layers. Each encoding layer consists of two 3D convolutional layers. Each 3D convolutional layer is followed by batch normalization and Relu activation function. At the end of each encoding layer, 3D maximum pooling operation is used to halve the resolution of the feature map. During the encoding process of the encoding layer, the number of feature channels of the second network unit is doubled. The bottleneck layer contains two consecutive 3D convolution operations; Each decoding layer contains a deconvolution operation and is used to upsample the feature map. The feature map formed after upsampling is concatenated with the corresponding encoding layer feature map through jump connections to combine low-level and high-level feature maps; the concatenated feature map undergoes two 3D convolution layers, batch normalization, and Relu activation function operations; a separate 3D convolution operation is added to the last decoding layer.

6. The method for generating a three-dimensional irregular elimination interface of a pigtail wishbone according to claim 5, characterized in that: There are three key positions of the second network unit, namely: the position after the Max pool in the third encoding layer, the position after the Up Conv in the first decoding layer, and the position after the Up Conv in the third decoding layer.

7. The method for generating a three-dimensional irregular elimination interface of a pigtail wishbone according to claim 5, characterized in that: The overall adaptive 3D segmentation network consists of 3 encoding layers, 1 bottleneck layer and 3 decoding layers; The first two coding layers of the overall adaptive three-dimensional segmentation network are the same as the first two coding layers in the second network unit. The third coding layer of the overall adaptive three-dimensional segmentation network includes the same structure as the third coding layer of the second network unit, and in the same structure as the third coding layer of the second network unit, the score in the complexity score processing layer is expanded to a tensor that is the same as the feature map generated after the Max pool operation. Then, the complexity score processing layer and the feature map generated after the Max pool operation are broadcasted, and the feature map generated after the broadcast operation is the deep feature map of the third coding layer. The broadcast operation introduces the key position after the Max pool in the third coding layer of the second network unit into the complexity score processing layer of this part. Finally, the deep feature map generated after the broadcast operation of the third coding layer is further selectively modeled through the Mamba module, and the feature expression ability of the network is enhanced through local convolution and state space mechanisms, so as to adapt to the feature requirements of different complexity areas, and form the third coding layer of the overall adaptive three-dimensional segmentation network. The bottleneck layer of the adaptive 3D segmentation network is used to connect the encoding layer and the decoding layer, and is the same as the bottleneck layer in the second network unit; The first decoding layer of the overall adaptive three-dimensional segmentation network includes the same structure as the first decoding layer of the second network unit. In the same structure as the first decoding layer of the second network unit, after deconvolution, the score in the complexity score processing layer will be expanded to the same tensor as the feature map generated after the deconvolution operation; then, the complexity score processing layer and the feature map generated after the deconvolution operation are broadcasted, and the feature map generated after the broadcasting operation is the deep feature map of the first decoding layer of the adaptive three-dimensional segmentation network. The broadcasting operation introduces the complexity score processing layer of this part after deconvolution in the first decoding layer of the second network unit; then, the deep feature map generated after the broadcasting operation of the first decoding layer is further selectively modeled through the Mamba module, and the feature expression ability of the network is enhanced through local convolution and state space mechanisms, so as to adapt to the feature requirements of different complexity areas; finally, after two 3D convolution layers, batch normalization and Relu activation function operations, the first decoding layer of the adaptive three-dimensional segmentation network is formed; The second decoding layer of the adaptive 3D segmentation network is the same as the second decoding layer in the second network unit; The third decoding layer of the overall adaptive three-dimensional segmentation network includes the same structure as the third decoding layer of the second network unit. In the same structure as the first decoding layer of the second network unit, the score in the complexity score processing layer will be expanded to the same tensor as the feature map generated after the deconvolution operation; then the complexity score processing layer and the feature map generated after the deconvolution operation are broadcasted, and the feature map generated after the broadcast operation is the deep feature map of the first decoding layer of the adaptive three-dimensional segmentation network. The broadcast operation introduces the complexity score processing layer of this part after the deconvolution in the first decoding layer of the second network unit; then, the Mamba module further selectively models the deep feature map generated after the broadcast operation of the first decoding layer, and enhances the feature expression ability of the network through local convolution and state space mechanisms, so as to adapt to the feature requirements of different complexity areas; finally, after two 3D convolution layers, batch normalization and Relu activation function operations, the third decoding layer of the adaptive three-dimensional segmentation network is formed.

8. The method for generating a three-dimensional irregular elimination interface of a pigtail wishbone according to claim 2, characterized in that: The adaptive 3D segmentation model is trained using the cross entropy loss function, an additional regularization term to balance the model, and a gradient descent optimization algorithm.

9. The method for generating a three-dimensional irregular elimination interface of a pigtail wishbone according to claim 1, characterized in that: Step 1 includes the following steps: Step 1.1: Based on the slice-by-slice images of the cross-section of the three-dimensional model of the pig hind leg, a threshold segmentation technique is used to obtain a preliminary segmentation line of the pig hind leg caudal wishbone; Step 1.2: Perform curvature analysis on the preliminary segmentation line to quantify the curvature of the bone in each slice image and obtain the curvature value of the preliminary segmentation line on each slice image; Step 1.3: Use the multi-directional slice curvature value data mapping method to construct a three-dimensional curvature model of the coccygeal wishbone.

10. The method for generating a three-dimensional irregular elimination interface of a pigtail wishbone according to claim 9, characterized in that: Step 2.1: Map the curvature data in each slice image in each direction to the three-dimensional space, obtain the three-dimensional point cloud containing curvature information in each direction, and set the number of block categories. , the 3D point cloud is divided into blocks; For a 3D point cloud, the feature vector of each point is composed of its 3D spatial position and curvature value The eigenvector is ; First, sort the curvature values ​​of all points. After sorting, the threshold can be determined directly based on the statistical distribution of the curvature values; after sorting the curvature values ​​from small to large, select multiple appropriate quantiles as the division The threshold of the complexity level area is set. According to the set threshold, the position of each point in the three-dimensional space of the coccyx is matched with its curvature category to form Point sets of different complexity, composed of Different block areas, i.e. blocks; Step 2.2: Use K-Means++ method to initialize cluster centers; A data point is randomly selected from the data set formed by the three-dimensional point cloud as the first cluster center; Calculate the Euclidean distance between each data point in the data set and the cluster center; for each data point that is not selected as a cluster center, calculate its distance to the nearest cluster center; assign the probability of being selected as the next cluster center based on the square of these distances, the greater the distance, the higher the probability of being selected; use the probability distribution calculated in the previous step to select a new cluster center by random sampling; repeat the above steps until a new cluster center is selected. The selected N cluster centers are the initial center points; Each time the K-Means++ algorithm selects a new center, it calculates the distance and probability based on the currently selected center and all data points. Suppose the coordinates of the data point P in the three-dimensional curvature model are , the coordinates of cluster center C are ; Step 2.3: Assign categories to the three-dimensional curvature model to obtain preliminary classification results; Assign each data point to the "cluster" where the nearest initial center point is located, forming Classification: Each part of the coccyx is classified into the group with the most similar complexity for preliminary classification; the category assignment process is as follows: for each data point , calculate its relationship with each cluster center Distance ,Will Assigned to the cluster center closest to it; Step 2.4: Update the cluster center, that is, according to the preliminary classification results, recalculate the center point of each cluster, that is, calculate the average position of all points in each category as the new cluster center; Step 2.5: Iterative optimization, that is, repeating steps 2.3 and 2.4, when the position of the cluster center tends to be stable or reaches the predetermined number of iterations, obtain The result of clusters is formed A sub-block composed of points of the same curvature category, that is, an area with similar complexity.

Images