CN114358388A - Method, device and equipment for planning surgical path for implanting pedicle screws - Google Patents

Abstract

The application provides a method, a device and equipment for planning a pedicle screw imbedding operation path, relates to the field of medicine, and can automatically plan the pedicle screw imbedding operation path and improve the accuracy and efficiency of pedicle screw imbedding operation path planning. The method comprises the following steps: carrying out vertebra example segmentation processing on a scanned image of a target vertebra to obtain image data of a plurality of different vertebra segments in the target vertebra; performing three-dimensional reconstruction on the image data of the plurality of vertebra segments to obtain three-dimensional vertebra images corresponding to the plurality of vertebra segments; according to the vertebra features, identifying and recording position information of key feature points corresponding to each vertebra segment in a scanned image; establishing a local coordinate system corresponding to the plurality of vertebral segments according to the position information; based on the plurality of three-dimensional vertebral images, locating a surgical path keypoint in a local coordinate system corresponding to the plurality of vertebral segments; and planning the surgical path of the corresponding vertebral segments according to the surgical path key points corresponding to the plurality of vertebral segments.

Description

Method, device and equipment for planning surgical path for implanting pedicle screws

Technical Field

The application relates to the field of medicine, in particular to a method, a device and equipment for planning a pedicle screw implantation operation path.

Background

The pedicle screw technology has good fixing effect on the aspect of spinal fixation, and the superiority of the pedicle screw technology enables the pedicle screw technology to be widely applied to spinal surgery. However, the existing preoperative screw position placing scheme is manually selected, and the operation is complex and time-consuming.

In an actual operation, the pedicle screw insertion point is determined by manually selecting the pedicle screw insertion point according to personal experience of a surgeon, so that the actual operation path is not consistent with the planned path, and the tail end instrument can not be accurately executed according to the planned path, so that the pedicle screw insertion effect is greatly differentiated. The problem to be solved at present is how to automatically plan the path for pedicle screw implantation, simplify the steps of path planning for pedicle screw implantation, and improve the efficiency and accuracy of path planning for pedicle screw implantation surgery.

Disclosure of Invention

The embodiment of the application provides a method, a device and equipment for planning a path for a pedicle screw imbedding operation, which can automatically plan the path for the pedicle screw imbedding operation, reduce the artificial dependence on surgeons in the pedicle screw imbedding operation, and improve the accuracy and efficiency of the path planning for the pedicle screw imbedding operation.

In a first aspect, the present application provides a method for planning a surgical path for pedicle screw implantation, comprising:

carrying out vertebra example segmentation processing on a scanned image of a target vertebra to obtain image data of a plurality of different vertebra segments in the target vertebra; performing three-dimensional reconstruction on the image data of the multiple vertebral segments to obtain three-dimensional vertebral images corresponding to the multiple vertebral segments; according to the vertebra features, identifying and recording position information of key feature points corresponding to each vertebra segment in the scanned image; establishing a local coordinate system corresponding to a plurality of the vertebra segments according to the position information; based on a plurality of the three-dimensional vertebral images, locating a surgical path keypoint in a local coordinate system corresponding to a plurality of the vertebral segments; and planning the surgical path of the corresponding vertebral segment according to the surgical path key points corresponding to the plurality of vertebral segments.

In the embodiment of the application, the image data of a plurality of different vertebra segments are obtained by segmenting the scanned image by using the vertebra example, the local coordinate system is established based on the position information of the key feature point of each vertebra segment obtained by the vertebra feature, and the operation path key point is positioned in the established local coordinate system by using the three-dimensional vertebra image reconstructed based on the image data of the plurality of vertebra segments, so that the operation path of the corresponding vertebra segment is planned according to the obtained operation path key point, thereby not only reducing the artificial dependence on a surgeon, but also improving the accuracy and the efficiency of the operation path planning of the pedicle screw implantation.

In a second aspect, the present application provides a pedicle screw placement surgical path planning device, comprising:

the example segmentation unit is used for carrying out vertebra example segmentation processing on the scanned image of the target vertebra to obtain image data of a plurality of different vertebra segments in the target vertebra;

the three-dimensional reconstruction unit is used for performing three-dimensional reconstruction on the image data of the multiple vertebral segments to obtain three-dimensional vertebral images corresponding to the multiple vertebral segments;

the position information recording unit is used for identifying and recording the position information of the key feature point corresponding to each vertebra segment in the scanned image according to the vertebra features;

the local coordinate system establishing unit is used for establishing a local coordinate system corresponding to the plurality of vertebra segments according to the position information;

a surgical path keypoint locating unit for locating a surgical path keypoint in a local coordinate system corresponding to a plurality of said vertebral segments based on a plurality of said three-dimensional vertebral images;

and the path planning unit is used for planning the operation path of the corresponding vertebral segment according to the operation path key points corresponding to the plurality of vertebral segments.

In a third aspect, the present application provides a pedicle screw placement surgical path planning device comprising a processor, a memory, and a computer program stored in the memory and executable on the processor, the processor implementing the method according to the first aspect or any of the alternatives of the first aspect when executing the computer program.

In a fourth aspect, the present application provides a computer readable storage medium storing a computer program which, when executed by a processor, implements a method according to the first aspect or any of the alternatives of the first aspect.

In a fifth aspect, the present application provides a computer program product, which when running on a pedicle screw placement operation path planning device, causes the pedicle screw placement operation path planning device to execute the steps of the pedicle screw placement operation path planning method according to the first aspect.

It is understood that the beneficial effects of the second aspect to the fifth aspect can be referred to the related description of the first aspect, and are not described herein again.

Drawings

In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings needed to be used in the embodiments or the prior art descriptions will be briefly described below, and it is obvious that the drawings in the following description are only some embodiments of the present application, and it is obvious for those skilled in the art to obtain other drawings based on these drawings without inventive exercise.

Fig. 1 is a schematic flowchart of a method for planning a surgical path for pedicle screw implantation according to an embodiment of the present disclosure;

FIG. 2 is a schematic diagram of a network architecture of a neural network model provided in an embodiment of the present application;

FIG. 3 is a schematic flow chart of a method of marking vertebral segments according to an embodiment of the present disclosure;

FIG. 4 is a set of labeling different vertebral levels with different labels provided by embodiments of the present application;

FIG. 5 is a schematic flow chart diagram illustrating a method of extracting image data of a vertebral segment according to an embodiment of the present disclosure;

fig. 6 is a schematic flowchart of a method for acquiring a first spine image according to an embodiment of the present disclosure;

fig. 7 is a set of image data in a vertebra example segmentation process provided in an embodiment of the present application, where (a) in fig. 7 is a scanned image of a target vertebra, fig. 7 (b) is image data obtained by performing gray-scale transformation and window adjustment on the image in (a) in fig. 7, (c) in fig. 7 is image data obtained by performing threshold division on the image in (b) in fig. 7, and fig. 7 (d) is image data obtained by screening a vertebra region in (c) in fig. 7;

FIG. 8 is a schematic diagram of a resampling provided by an embodiment of the present application;

fig. 9 (a) is a first spine image provided in the embodiment of the present application, and fig. 9 (b) is a three-dimensional vertebra image of a target spine provided in the embodiment of the present application;

fig. 10 is a schematic flowchart of a method for acquiring location information of a key feature point according to an embodiment of the present application;

fig. 11 (a) is an image of a closed spinal region, fig. 11 (b) is an image of an open spinal region, and fig. 11 (c) is an image of a spinal region having a wrong judgment;

FIG. 12 is a flowchart illustrating a method for determining a reference plane of a local coordinate system according to an embodiment of the present application;

FIG. 13 is a schematic view of an image of a vertebral axis provided by an embodiment of the present application;

fig. 14 is a schematic image of a second reference plane S provided in the embodiment of the present application;

FIG. 15 is a schematic diagram of an image depicting a three-dimensional vertebral image placed in its corresponding local coordinate system according to an embodiment of the present application;

fig. 16 (a) is a three-dimensional vertebral image before an uncorrected orientation according to an embodiment of the present application, and fig. 16 (b) is a three-dimensional vertebral image after a corrected orientation according to an embodiment of the present application;

FIG. 17 is a flowchart illustrating a method for locating a surgical path keypoint in accordance with an embodiment of the present disclosure;

FIG. 18 is an image schematic of an embodiment of the present application providing an image on a three-dimensional vertebral image taken in a plurality of cross-sections containing the shape of the pedicle region;

FIG. 19 is a schematic representation of an image of a bilateral pedicle region as fully encompassed within the demarcated areas as contemplated by the practice of the present application;

FIG. 20 is an image of a surgical path keypoint provided by an embodiment of the present application;

fig. 21 is a diagram of a path planning result provided in the embodiment of the present application;

fig. 22 is a schematic structural diagram of a pedicle screw placement surgical path planning device according to an embodiment of the present application;

fig. 23 is a schematic structural diagram of a pedicle screw placement surgical path planning device according to an embodiment of the present application.

Detailed Description

In the following description, for purposes of explanation and not limitation, specific details are set forth, such as particular system structures, techniques, etc. in order to provide a thorough understanding of the embodiments of the present application. It will be apparent, however, to one skilled in the art that the present application may be practiced in other embodiments that depart from these specific details. In other instances, detailed descriptions of well-known systems, devices, circuits, and methods are omitted so as not to obscure the description of the present application with unnecessary detail.

It should be understood that the term "and/or" as used in this specification and the appended claims refers to and includes any and all possible combinations of one or more of the associated listed items. Furthermore, in the description of the present application and the appended claims, the terms "first," "second," "third," and the like are used for distinguishing between descriptions and not necessarily for describing or implying relative importance.

It should also be appreciated that reference throughout this specification to "one embodiment" or "some embodiments," or the like, means that a particular feature, structure, or characteristic described in connection with the embodiment is included in one or more embodiments of the present application. Thus, appearances of the phrases "in one embodiment," "in some embodiments," "in other embodiments," or the like, in various places throughout this specification are not necessarily all referring to the same embodiment, but rather "one or more but not all embodiments" unless specifically stated otherwise. The terms "comprising," "including," "having," and variations thereof mean "including, but not limited to," unless expressly specified otherwise.

The application provides a method for planning a path for a pedicle screw implantation operation, which comprises the steps of firstly obtaining a CT image before a patient operation, namely a scanning image in the following text, completing segmentation of a vertebral segment by using a deep learning neural network model to obtain a plurality of different vertebral segments, then carrying out feature recognition and extraction according to image data of different vertebral segments, establishing a local coordinate system of each vertebral segment, and automatically planning an operation path of each vertebral segment in the pedicle screw implantation operation on the local coordinate system. Through introducing the neural network model of degree of depth study for can be fast convenient the segmentation go out all kinds of complicated scanning images, compare with traditional manual selection and simplified the flow greatly, reduced the time in the operation process and improved the commonality, not only reduced the artificial dependence to surgical doctor, improved pedicle of vertebral arch screw moreover and put into operation path planning's rate of accuracy and efficiency. By evaluating the path planning result, the yield of 93% is obtained, and the effectiveness of the extracted pedicle screw implantation operation path planning method is verified.

Referring to fig. 1, fig. 1 is a schematic flow chart of a method for planning a pedicle screw placement operation path according to an embodiment of the present application, which is detailed as follows:

step S101, vertebra example segmentation processing is carried out on the scanned image of the target vertebra, and image data of a plurality of different vertebra segments in the target vertebra are obtained.

In the embodiment of the present application, the scan image is a two-dimensional plane scan image obtained by performing a plane scan on the target vertebra, including but not limited to a scan image obtained by means of a CT imaging device. Before an operation, CT scanning is carried out on a patient through a CT imaging device, and a scanning image of a target vertebra is obtained.

After the scanned image of the target vertebra is obtained, vertebra example segmentation processing is carried out on the scanned image through a trained neural network model to obtain each vertebra segment through segmentation, namely segmentation and extraction of different examples of each vertebra segment are completed from the scanned image of the whole section of the target vertebra.

Referring to fig. 2, fig. 2 is a schematic diagram of a network architecture of a neural network model according to an embodiment of the present disclosure.

As shown in fig. 2, the network architecture of the neural network model is a full convolution iterative neural network architecture, and mainly comprises three parts, wherein the main part, i.e. the first part, is a segmentation network for segmenting different vertebra segments from the scanned image of the target vertebra; the second part is an example storage memory module which is used for marking the divided vertebra segments and simultaneously inputting the marked vertebra segments with different labels and the data set into a division network, so that the process of dividing the vertebra segments is from top to bottom, and the division network is triggered to focus on the next vertebra segment in the next iteration process; the third part is a classification network for distinguishing different vertebral regions.

It should be noted that the data set in the above is spine image data used for training the neural network model to obtain good parameters to achieve the segmentation objective. For example, the data set may be the Computerized Spine Imaging (CSI) from MICCAI 2014, which is mainly composed of CT scan images of the human Spine, including image data of all thoracic and lumbar vertebra segments and different vertebral segments labeled with different labels, the scan subject being a healthy young year of 20-34 years old.

When the neural network model is trained, the data set is divided into two groups, namely a training set and a testing level, wherein the image data of the training set is used for training the neural network model, the image data in the testing set is used for verifying the training effect of the neural network model, the error of the trained neural network model on the testing set is calculated, and the neural network model with the minimum error of the trained neural network model on the testing set is used as the final trained neural network model.

As shown in fig. 2, the numbers above the tile represent the number of channels of the feature image, and the numbers to the left of the tile represent the size of the feature image. The input data of the input layer in the neural network model is a 128 × 128 × 128 voxel block containing 2 data channels, which is formed by splicing an input data block and an example storage data block, wherein the input data block is image data containing a vertebra, such as original vertebra scanning image data used for training, scanning image data of a target vertebra and the like, and the example storage data block is image data of a vertebra segment with a label stored in an example storage memory module. The input data is convolved to form a 128 x 128 voxel block comprising 84 data channels. The voxel block is three-dimensional volume data formed by splicing a plurality of groups of scanning images, and the size of the three-dimensional volume data is 128 multiplied by 128.

In the neural network model, the data input into the neural network model are a plurality of or all of the scanned images with labels in the data set, the scanned image of the target vertebra and the vertebrae example segmentation data obtained by the last training.

As shown in fig. 2, after the input image data is operated by the neural network model, a one-dimensional feature vector is obtained by performing a densen operation on the last layer, and the one-dimensional feature vector is probability prediction that a vertebra is completely visible in a process of vertebrae example segmentation, and is mainly used for judging the integrity of the vertebra. Complete visibility evaluation on the vertebrae is completed by using the one-dimensional feature vector, a final vertebrae instance segmentation task is completed by using an instance storage memory module and a classification network, and different labels of different vertebrae are extracted.

The hidden layer in the neural network model comprises convolution operation, normalization operation, excitation operation and upsampling. In each hidden layer, convolution is completed by utilizing two times of 3 multiplied by 3 image blocks, and pooling operation is added after convolution, so that invariance of data is guaranteed, parameters and calculated amount are reduced while characteristics are kept, dimension reduction is realized, overfitting is prevented, and generalization capability is improved. Meanwhile, a ReLu activation function is adopted in each hidden layer to add a nonlinear unit, so that fitting is reduced, more efficient gradient descent and backward propagation are generated, and the problems of gradient explosion and gradient disappearance are avoided. And the depth of the network is continuously increased by multi-level convolution and pooling operation, and deeper scale information contained in the image is mined. The weight and the bias in the convolution process are continuously adjusted in an iterative manner by the whole neural network model during back propagation so as to achieve a smaller loss function and realize the optimization of the classification effect.

In the convolution operation process, an upsampling branch is added, an intermediate instance segmentation storage module with the size similar to that of the input data set is obtained in an intermediate link such as an instance storage memory module through the upsampling process, and the currently completed instance segmentation is marked in the instance storage memory module, so that the currently completed instance segmentation is used as input for the first time, and guidance is provided for the next iteration segmentation process. And the other branch continuously convolving downwards is used as an output layer of the neural network model, a one-dimensional vector is output, example classification and division are completed by using Sigmoid excitation function operation, and the vertebra integrity score is output. Where the one-dimensional vector is used to assess the integrity of the vertebral level, i.e. the probability prediction that the vertebra is fully visible during example segmentation. In the classification network, the segmented vertebral segments are labeled with different labels according to the vertebral characteristics. In the process of training the neural network model, iterative optimization is carried out on the neural network model through a loss function, so that the error of the trained neural network model on a test set is minimized, and the example segmentation accuracy is improved.

In the embodiment of the present application, the calculation of the loss function includes a segmentation error and a classification error, where the loss function is specifically as follows:

e_segment＝λ·FP_soft+FN_soft；

e_classify＝-tlogp-(1-t)log(1-p)；

wherein the content of the first and second substances,

as a loss function, e_segmentIs a segmentation error; e.g. of the type_classifyIs a classification error; FP_softFalse positive rate; FN (FN)_softFalse negative rate; λ is the set weight; t is a truth label (0 or 1), and p is a prediction probability value; t is t_iIs a binary truth label, p_iIs t_iCorresponding probability predicted values; omega_iIs a weight coefficient, and gamma and sigma are adjustable parameters; d_iRepresenting the distance of the current voxel point position in the three-dimensional volume data from the vertebra.

In the embodiment of the present application, when the neural network model is trained, λ is 0.1, γ is 8, and σ is 6. When the neural network model is realized by using a pytorre framework, the learning rate is set to be 0.001, the dynamic increment is set to be 0.99, and the iteration number is 40000 times. In the iterative process, the training parameters of the model are recorded when the loss function is subjected to reduction transformation, and finally the minimum ratio of the loss function such as

The neural network model has good segmentation effect on the spine, and lays a good foundation for planning the operation path for placing the pedicle screws.

Referring to fig. 3, fig. 3 is a schematic flow chart of a method for marking a vertebral segment according to an embodiment of the present application, which is detailed as follows:

and S301, performing vertebra example segmentation processing on the scanned image through the trained neural network model.

In the embodiment of the application, the scanned image and the data set are input into a trained neural network model, and after convolution operation, normalization operation, excitation operation and up-sampling processing are carried out on the scanned image and the data set through a segmentation network, an example storage memory module and a classification network of the neural network model, a certain example, namely a certain vertebral segment in a target vertebra, is obtained through segmentation.

Step S302, according to the vertebra characteristics, a label corresponding to the vertebra segment obtained by current segmentation is searched.

In the embodiment of the application, according to the vertebra features in the data set, a label corresponding to a currently segmented vertebra segment is searched, and different vertebra features are marked by using different labels, for example, different vertebra features are marked by using different colors, characters, color-character combinations and the like.

Step S303, marking the currently segmented vertebra segment by using the label.

In the embodiment of the application, after the label corresponding to the currently segmented vertebral segment is found, the label is used for marking the currently segmented vertebral segment, and the labeled currently segmented vertebral segment is used as a parameter for iterative optimization of the neural network model.

And step S304, performing iterative optimization processing by taking the marked vertebra segment as the input of the neural network model, and outputting the next vertebra segment until image data of a plurality of different vertebra segments in the target vertebra are obtained.

In the embodiment of the application, after the corresponding label of the vertebra segment obtained by current segmentation is found, the label is used for marking the vertebra segment, the marked vertebra segment is input into the neural network model again for iterative optimization processing, and the next vertebra segment is guided and output until the image data of a plurality of different vertebra segments in the target vertebra are obtained.

It should be noted that the data set includes spine image data of different integrity degrees, such as different types of image data of a whole spine image, a partial spine image, and the like, so that the scanned image with high complexity and high uncertainty still has a good segmentation effect, and the segmentation of different vertebra segments can be completed according to the result of the label marking, so that each segment of vertebra data can be accurately and quickly extracted.

In some embodiments of the present application, a set of scanned images taken at different angles or at different orientations are subjected to a vertebral instance segmentation process by a neural network, resulting in a set of labeled different vertebral segments using different labels as shown in fig. 4. As shown in FIG. 4, each vertebral level is marked with a different color, such as from top to bottom, a first vertebral level with red color and a second vertebral level with orange color … …

Referring to fig. 5, fig. 5 is a schematic flowchart illustrating a method for extracting image data of a vertebral segment according to an embodiment of the present application, which is detailed as follows:

step S501, removing noise in the scanned image to obtain a first spine image, where the noise is image data of the scanned image except for the vertebra segment.

In the embodiment of the application, due to the existence of each complex structure in a human body and the problem of image shooting quality, the obtained scanning image includes more image noise and the existence of bones in other areas in the same cross section, such as ribs, and in order to provide the accuracy and efficiency of vertebral segment segmentation, the image data needs to be removed to obtain the image data of the vertebral region of the target vertebra, and then the image data of the vertebral region is subjected to example segmentation processing through a neural network.

Referring to fig. 6, fig. 6 is a schematic flowchart of a method for acquiring a first spine image according to an embodiment of the present application, which is detailed as follows:

step S601, performing image preprocessing on the scanned image to obtain a third spine image, where the image preprocessing includes gray scale conversion, window adjustment and threshold division.

In the embodiment of the present application, a third spine image is obtained by performing gray scale conversion, window adjustment processing, and threshold division processing on a scanned image. The third spine image is image data obtained by removing a soft tissue portion in the scanned image, that is, the third spine image is image data including a bone portion, including but not limited to image data of bone portions such as spine, rib, sternum, and the like.

Step S602, calculating an area of a connected region existing in the third spine image to obtain an area of at least one connected region.

Because the third spine image further includes image data of other bones besides vertebrae, in order to improve the example segmentation effect of the vertebral segment, a vertebra region needs to be screened out from the third spine image to reduce the image of the image data of other bones for carrying out vertebra example segmentation on the neural network model, so that the accuracy of vertebra example segmentation is improved.

In the embodiment of the application, integral calculation is performed through the connected region existing in the third spine image, the area of the connected region existing in the third spine image is calculated from the region with the similar size of the spine region, namely the spine region, so as to obtain the area of at least one connected region, and the vertebra part required by path planning can be effectively reserved, so that the accuracy of the path planning is improved.

In some embodiments of the present application, the area of the connected region is calculated by traversing the third spine image through a connected region area extraction calculation formula, where the connected region area extraction calculation formula is:

Do B_i+1＝(B_i⊕S)∩A

Until B_i+1＝＝B_i

wherein A is a communication region, B_iRepresenting a certain pixel point in the communication area A, wherein i is an integer and is more than or equal to 0; the operators used for the representation of values and operations are in the form of a 3 x 3 squared matrix, filled with the number 1.

It should be noted that, when the area of the communication area is calculated, an iterative loop is performed according to the communication area extraction calculation formula until B_iAnd filling the whole communication area A to finish the extraction of the communication area A.

Step S603, determining a vertebra region in the third vertebra image according to at least one of the communication area areas.

In this embodiment, when a plurality of connected regions exist in the third spine image, a plurality of connected region areas are obtained, and the obtained area sizes of the plurality of connected regions are compared to determine the vertebra region in the third spine image. For example, a vertebra region in the third vertebra image is determined according to a first large communication region, wherein the first large communication region is a communication region corresponding to the largest area in the communication region areas.

Step S604, extracting image data corresponding to the vertebra region from the third vertebra image, to obtain the first vertebra image.

In this embodiment, after the vertebra region in the third vertebra image is determined by following the first large connected region, the image data in the first large connected region is extracted to obtain the first vertebra image, that is, the image only including the vertebra segment.

As shown in fig. 7, fig. 7 is a set of image data in a vertebra example segmentation process provided in an embodiment of the present application, where (a) in fig. 7 is a scanned image of a target vertebra, fig. 7 (b) is image data obtained by performing gray scale transformation and window adjustment processing on the image of (a) in fig. 7, (c) in fig. 7 is image data obtained by performing threshold division processing on the image of (b) in fig. 7, and fig. 7 (d) is image data obtained by performing screening on a vertebra region of (c) in fig. 7.

Step S502, resampling image data in the first spine image to obtain a second spine image in a predetermined resolution direction.

In the embodiment of the present application, due to the reason of the CT device, in most cases, the pixel pitch in the scanned image is not matched with the layer pitch, so that after the first spine image is extracted, when three-dimensional volume data is to be reconstructed from the scanned image, for example, three-dimensional volume data corresponding to a vertebra segment is obtained based on image data of the vertebra segment through three-dimensional reconstruction, a large deviation occurs, which causes a large error to occur in a finally planned surgical path, and in order to avoid this situation, it is necessary to resample the obtained first spine image to complete an interpolation and completion process.

Because the pixel pitch and the layer pitch of the CT scanning image during tomography are different, the image data in the first spine image needs to be resampled to a predetermined resolution direction, such as a low resolution direction, by a resampling calculation formula to obtain a second spine image when the CT scanning image is converted into three-dimensional volume data.

In the embodiment of the present application, the resampling calculation formula is specifically as follows:

A_p＝1-p-(1-2p)·D_a

A_q＝1-q-(1-2q)·D_b

A_r＝1-r-(1-2r)·D_c

wherein model (i, j, k) is the interpolated point, and (i, j, k) is the coordinate information of the interpolated point; dcm (x + p, y + q, z + r) is 8 vertexes of an adjacent cube taking an interpolation point as a center, (x, y, z) is coordinate information of a certain pixel point, p, q, r are auxiliary parameters, and take values of 0 and 1 respectively to represent serial numbers of 8 fixed points in the adjacent cube (as shown in fig. 8, fig. 8 is a resampling schematic diagram provided by the embodiment of the present application); a. the_p,A_q,A_rFor auxiliary multipliers in the resampling calculation process, D_a,D_bThe pixel pitch in the cross section of the CT scan is shown. D_cIs the layer spacing in CT scanning.

Step S503, extracting image data of a corresponding vertebra segment from the second vertebra image according to the labeled label corresponding to the vertebra instance.

In the embodiment of the application, through a resampling calculation formula, missing points, namely interpolated points, when the first spine image is converted into the three-dimensional second spine image can be calculated, and then according to the labeled labels corresponding to the vertebra examples, image data of the vertebra segments corresponding to the vertebra examples can be extracted from the second spine image, so that image data of a plurality of different vertebra segments can be obtained, and therefore three-dimensional reconstruction is performed on the basis of the extracted image data of the plurality of vertebra segments, and three-dimensional data which is more accurate and has practical significance can be obtained.

It should be noted that the obtained image data of a plurality of different vertebral segments includes interpolated points calculated by a resampling calculation formula, so that it is ensured that a three-dimensional vertebral image with higher accuracy can be obtained when three-dimensional reconstruction is performed based on the image data.

And step S102, performing three-dimensional reconstruction on the image data of the plurality of vertebra segments to obtain three-dimensional vertebra images corresponding to the plurality of vertebra segments.

As shown in fig. 9, (a) in fig. 9 is a first spine image provided by the embodiment of the present application, and (b) in fig. 9 is a three-dimensional vertebra image of a target spine provided by the embodiment of the present application, it can be seen from the figure that, after resampling the first spine image by the resampling calculation formula, more image data and more real three-dimensional volume data are obtained.

And step S103, identifying and recording the position information of the key feature point corresponding to each vertebral level in the scanned image according to the vertebral features.

In the embodiment of the application, the key characteristic points exist on each vertebral segment, the local coordinate system corresponding to the vertebral segment can be established by determining the position information of the key characteristic points, the specific analysis can be carried out by the local coordinate systems corresponding to different vertebral segments, the related nail path information for positioning the pedicle of vertebral arch can be determined, and therefore a more accurate operation path is constructed.

The vertebra features, which may also be referred to as anatomical shape features, may identify key feature points from a three-dimensional vertebra image of a target vertebra according to the vertebra features, thereby obtaining position information of the identified key feature points, and completing a local coordinate system of a corresponding vertebra segment based on the position information.

And step S104, establishing a plurality of local coordinate systems corresponding to the vertebra segments according to the position information.

In the embodiment of the application, a local coordinate system corresponding to each vertebral segment is established according to the position information of the key feature points of each vertebral segment. The position information comprises position information of a central point of the vertebral canal and position information of a central point of the vertebral body, namely the position information comprises the position information of the central point of the vertebral canal and the position information of the central point of the vertebral body.

Referring to fig. 10, fig. 10 is a schematic flowchart of a method for acquiring location information of a key feature point according to an embodiment of the present application, which is detailed as follows:

and S1001, cutting out a vertebra cross section image from the scanned image through a preset cross section.

In the embodiment of the present application, the preset cross section is a plane with different shapes, including but not limited to a plane with a square shape, a rectangle shape, a diamond shape, etc., and the scanned image is sliced through the preset cross section, so that a plurality of cross-sectional images of the vertebra are obtained by cutting out from the scanned image, and in each cross-sectional image of the vertebra, the main image information to be identified includes a spinal canal region and a vertebral body region.

Step S1002, searching a spinal canal region to be identified from the cross-sectional image of the vertebra according to a second large communication area, where the second large communication area is a communication area corresponding to a second large area of the communication area.

In this application embodiment, because in continuous many vertebrae cross sectional images, the shape of each canalis spinalis is different, the regional condition of closed canalis spinalis probably exists, thereby lead to the problem of wrong judgement canalis spinalis center to appear, in order to avoid this condition to appear, the robustness of the algorithm of reinforcing solution canalis spinalis center makes it can accurately seek canalis spinalis center, need confirm the canalis spinalis region according to the big connected area of second after, count up the data of the regional area of canalis spinalis, the canalis spinalis center of guaranteeing to discern through calculation and screening is in accurate position.

And eliminating data with overlarge area (the vertebral canal is not closed) and undersized area (the vertebral canal is judged wrongly) in the statistical result, drawing a frequency distribution map according to the result, and selecting an interval with highest frequency, namely an interval of the vertebral segment with the closed vertebral canal and the most clear shape, so that the center of the vertebral canal is recognized on the vertebral bodies of different shapes and different sizes of each vertebral segment, and the algorithm has stability.

As shown in fig. 11, (a) in fig. 11 is an image of a closed spinal canal region, (b) in fig. 11 is an image of an unclosed spinal canal region, and (c) in fig. 11 is an image of a spinal canal region where there is a wrong judgment, which is due to the presence of an unclosed spinal canal region and another closed non-spinal canal, resulting in a wrong judgment.

Step S1003, performing area integration and center solution on the vertebral canal region to obtain a vertebral canal center point P in the vertebral cross section image_i。

In the embodiment of the application, the integral averaging is carried out on the coordinates x and y of each pixel point in the vertebral canal region, so that the vertebral canal region is solvedThe center of mass of the vertebral canal can be obtained, namely the central point P of the vertebral canal in the cross section image of the vertebra_i。

Step S1004, obtaining a centrum center point Q of a centrum region in the transverse section image of the vertebra through a maximum inscribed circle geometric calculation method_i。

In the embodiment of the application, for each vertebra cross section image, a centrum center point Q of a centrum region in the vertebra cross section image is obtained by adopting a Voronoi diagram to calculate a maximum inscribed circle geometric calculation method_i。

Before establishing the local coordinate system of each vertebral level, it is necessary to determine the origin of the local coordinate system and three reference planes, which are named first, second and third reference planes in turn, corresponding to the three planes between the x and y, x and z, y and z axes of the coordinate system.

A plurality of vertebral cross-sectional images are obtained by intercepting the scanned image through a series of vertebral cross-sectional images corresponding to the same vertebral segment, namely through a first preset cross section, so that a plurality of vertebral canal central points, a plurality of vertebral body central points and position information of each vertebral canal central point and each vertebral body central point can be obtained, and after the position information of the vertebral canal central point serving as the origin of a local coordinate system is determined, a reference plane of the local coordinate system is determined according to the position information of the vertebral canal central point and the position information of the vertebral body central point; and establishing a local coordinate system corresponding to the vertebra segments according to the origin and the reference plane.

Referring to fig. 12, fig. 12 is a flowchart illustrating a method for determining a reference plane of a local coordinate system according to an embodiment of the present application, which is detailed as follows:

and step S1201, determining the axis of the vertebra according to the position information of the central point of the vertebral canal by using a least square fitting linear formula of the coordinate points in the space.

In this embodiment of the application, through the first preset cross section, a plurality of cross-sectional images of vertebrae are obtained by cutting from the scanned image, and then based on the obtained cross-sectional images of vertebrae, position information of a plurality of central points of vertebral canal and a plurality of central points of vertebral body is obtained, and the axis of vertebrae as shown in fig. 13 can be calculated by using a space coordinate point least square fitting straight line formula, wherein the space coordinate point least square fitting straight line formula is as follows:

x＝a·z+b

y＝c·z+d

wherein a, b, c and d are parameters of a linear equation in space, and x, y and z are points on the linear equation in space and are used for representing a straight line in space; p_ix,P_iy,P_izIs the central point P of the vertebral canal_iN is the central point P of the vertebral canal_iThe number of the cells.

Step S1202, a plane perpendicular to the vertebral axis is taken as a first reference plane of the local coordinate system.

In the embodiment of the application, after the vertebra axis is determined, a plane perpendicular to the vertebra axis is made, and the plane perpendicular to the vertebra axis is a first reference plane of the local coordinate system.

Step S1203, fitting a second reference plane meeting a preset condition according to the first parameter of the plane system equation meeting the optimization condition, where the preset condition is a distance condition between the vertebral axis and the position information of the vertebral body center point.

In the embodiment of the present application, before determining the second reference plane, it is necessary to calculate that the second reference plane maximally approaches all vertebral body center points Q_iThat is, the shortest distance between the center point of the vertebral body and the plane passing through the vertebral axis plane system in the space needs to be calculated, and the sum of the distances is taken as an optimization target to enable the second reference plane to be close to all the center points Q of the vertebral body to the maximum extent_i。

Here, the first parameter is calculated by a first parameter calculation formula, wherein the first parameter calculation formula is specifically as follows:

x-a·z-b+k·(y-c·z-d)＝0

wherein a, b, c and d are parameters of a linear equation in space, k is a parameter of a plane system equation, and Q_ix,Q_iy,Q_izIs the central point Q of the vertebral body_iThe coordinate information of (2).

Calculating to obtain a first parameter k of a plane system equation satisfying the optimization condition, and fitting to obtain a plane system equation passing through the axis of the vertebra and closest to the central points Q of all the vertebral bodies as shown in FIG. 14_iAs a plane of symmetry of the current vertebral level, i.e. the second reference plane S.

Step S1204, determining a third reference plane of the local coordinate system according to the first reference plane and the second reference plane.

In the embodiment of the present application, after the first reference plane and the second reference plane are obtained, the third reference plane in the spatial coordinate system may be directly calculated according to the first reference plane and the second reference plane. After the origin of the spatial coordinate system is determined, a normalized local coordinate system can be obtained. The vertebral foramen center closest to the center of the whole vertebral body segment is selected as the origin of the space coordinate system, and the local coordinate system corresponding to the vertebral body segment can be obtained.

And step S105, positioning a key point of an operation path in a local coordinate system corresponding to a plurality of vertebra segments on the basis of a plurality of three-dimensional vertebra images.

In the embodiment of the application, the key points of the operation path are the path points of the nail path for implanting the pedicle screws. As shown in fig. 15, each three-dimensional vertebra image is described in its corresponding local coordinate system, which facilitates subsequent accurate surgical path planning based on the identification of vertebra features and vertebra regions.

When the scanned image of the target vertebra is subjected to vertebra example segmentation processing, the obtained multiple vertebra segments keep the original postures, so that the vertebra poses in the finally obtained three-dimensional vertebra image also keep the original postures, which is not in accordance with the conventional observation habit and is inconvenient for calculation and planning of an operation path.

In some embodiments of the present application, the orientation correction may be performed on the vertebral pose of the three-dimensional vertebral image before the uncorrected orientation shown in (a) in fig. 16 by rotating the matrix Rver, resulting in a corrected backward three-dimensional vertebral image as shown in (b) in fig. 16.

Referring to fig. 17, fig. 17 is a schematic flow chart of a method for positioning a surgical path keypoint according to an embodiment of the present disclosure, which is detailed as follows:

step S1701, a first cross section of the three-dimensional vertebral image is taken through a vertebral assistant cross section, the first cross section being a plurality of cross sections taken from corresponding vertebral levels along a predetermined direction of the vertebral assistant cross section.

In an embodiment of the application, the vertebral auxiliary section is a plane perpendicular to the second reference plane. The predetermined direction is defined by the central point P of the vertebral canal_iAnd the centrum center point Q_iIn the axial direction of the line, i.e. with the auxiliary section of the vertebra in line P_iQ_iAnd (4) moving upwards. As shown in fig. 18, through the auxiliary section at line P_iQ_iThe upper movement can intuitively obtain a plurality of sections containing the shape of the pedicle area.

Step 1702, by a region segmentation method, obtaining a set of cross-sections by segmenting the first cross-section according to a first region and a second region, where the set of cross-sections includes a second cross-section and a third cross-section.

In the embodiment of the application, the first cross section is divided into two parts according to the left area and the right area by using a region division algorithm to obtain two cross sections.

As shown in fig. 19, for stability of the region segmentation algorithm, when performing region segmentation, a margin of a certain threshold value needs to be reserved in both regions, so that both pedicle regions can be completely included in the segmented region, and further, information point data of both regions are calculated in the first region and the second region, respectively.

Step S1703, determining a set of center position points corresponding to the set of cross sections by using a method of drawing a maximum inscribed circle.

In the embodiment of the present application, the central position point of the pedicle region of the first cross section and the central position point of the pedicle region of the second cross section is found by continuously drawing the maximum inscribed circle on the cross section.

Step S1704, a group of pedicle center position points is determined according to the plurality of groups of center position points, wherein the group of pedicle center position points comprises a first pedicle center position point and a second pedicle center position point.

In the embodiment of the application, the central position points of the multiple groups of sections can be obtained by adopting a method of drawing the maximum inscribed circle. Then, the center position points of the pedicle regions of the plurality of sets of cross sections are divided into two groups according to the region division, for example, the center position points of the pedicle regions in the cross sections divided by the first region (for example, the left region of the first cross section) in the plurality of sets of cross sections are divided into one group, and the center position points of the pedicle regions in the cross sections divided by the second region (for example, the right region of the first cross section) in the plurality of sets of cross sections are divided into another group, so that two updated groups of center position points are obtained.

Here, the first pedicle center position point is an average value of the sum of center position points determined for the section located in the first region, and the second pedicle center position point is an average value of the sum of center position points determined for the section located in the second region.

In some embodiments of the present application, the two sets of center position points are solved by a pedicle center position point solving formula to obtain center position points of the pedicle region of two regions, such as a first region located on the left side and a second region located on the right side, and the pedicle center position point solving formula is as follows:

wherein C is a central position point of the pedicle region in the first or second region, C_iThe center of the maximum circumscribed circle obtained on the second section or the third section, and m is the number of the first sections.

And step S1705, calculating the path intersection point of the pedicle screw implantation operation through a pedicle intersection point calculation formula.

In the embodiment of the present application, according to the anatomy of human vertebra, it can be known that, due to the symmetry characteristics of human body, the intersection point of the straight lines of the left and right pedicle path channels is usually located on the symmetry axis of the end of vertebra, the intersection point M of the pedicle paths can be determined by different sizes of the vertebra segments, and the intersection point of the pedicle paths on both sides is calculated by the formula of the intersection point of the pedicle paths, wherein the formula of the intersection point of the pedicle paths is as follows: is composed of

QM＝t·PQ

Wherein M is the intersection point of the pedicle paths; p is the central point P of the vertebral canal_iThe central point of a vertebral canal closest to the origin in the local coordinate system, and Q is the central point Q of the vertebral body_iA centrum center point of the centrum closest to the origin in the local coordinate system; t is a proportionality coefficient, different vertebral segments correspond to different proportionality coefficients t, the proportionality coefficient t is a value selected empirically after a plurality of tests, and can be specifically adjusted according to experimental conditions, for example, the proportionality coefficient t in the lumbar vertebra of the segment L1-L4 is 1, and the proportionality coefficient t in the segment L5 is 0.7.

And step S1706, determining a key point of the operation path according to the group of the central position points of the vertebral pedicle and the path intersection point.

In the examples of the present application, C is used_leftRepresenting the first pedicle center location point, by C_rightRepresenting the first pedicle center location point.

Referring to fig. 20, fig. 20 is a schematic view of an image of a surgical path key point according to an embodiment of the present application; as shown in fig. 20, the first pedicle center location point C is located_leftConnecting with the path intersection point M to obtain a first path planning straight line MC_leftThe center of the second vertebral pedicle is positioned at a point C_rightConnecting with the path intersection point M to obtain a second pathRadial planning straight line MC_rightWherein the first path defines a straight line MC_leftDrawing a straight line MC with a second path_rightIntersect at path intersection point M.

Obtaining a first path planning straight line MC from a three-dimensional vertebral image_leftAnd a second path planning straight line MC_rightAnd recording the coordinate point information, wherein the obtained coordinate point information is the key point of the operation path.

And S106, planning the operation path of the corresponding vertebral segment according to the operation path key points corresponding to the plurality of vertebral segments.

In the embodiment of the application, after the operation path key point is determined, a corresponding space linear equation can be obtained according to the coordinate information of the operation path key point, and the space linear equation is a nail path planning path of the current vertebral segment, namely an operation path for pedicle screw placement.

Here, the planned surgical path is displayed in a three-dimensional vertebral image, resulting in a path planning result map as shown in fig. 21. It can be seen from the path planning result graph that the planned surgical path is located at the center of the pedicle region, and a good surgical access channel is drawn by placing a pedicle screw using a CT image as input data into the surgical path planning, so that the automatic planning of the nail path for a single vertebra is completed.

In the embodiment of the application, the image data of a plurality of different vertebra segments are obtained by segmenting the scanned image by using the vertebra example, the local coordinate system is established based on the position information of the key feature point of each vertebra segment obtained by the vertebra feature, and the operation path key point is positioned in the established local coordinate system by using the three-dimensional vertebra image reconstructed based on the image data of the plurality of vertebra segments, so that the operation path of the corresponding vertebra segment is planned according to the obtained operation path key point, thereby not only reducing the artificial dependence on a surgeon, but also improving the accuracy and the efficiency of the operation path planning of the pedicle screw implantation.

It should be understood that, the sequence numbers of the steps in the foregoing embodiments do not imply an execution sequence, and the execution sequence of each process should be determined by its function and inherent logic, and should not constitute any limitation to the implementation process of the embodiments of the present application.

Based on the pedicle screw implantation surgical path planning method provided by the embodiment, the embodiment of the application further provides an embodiment of a device for implementing the embodiment of the method.

Referring to fig. 22, fig. 22 is a schematic view of a pedicle screw placement surgical path planning device according to an embodiment of the present application. The units are included for performing the steps in the corresponding embodiment of fig. 1. Please refer to fig. 1 for the related description of the corresponding embodiment. For convenience of explanation, only the portions related to the present embodiment are shown.

Referring to fig. 22, the pedicle screw placement operation path planning device 22 includes:

an example segmentation unit 221, configured to perform vertebra example segmentation processing on a scanned image of a target vertebra to obtain image data of a plurality of different vertebra segments in the target vertebra;

a three-dimensional reconstruction unit 222, configured to perform three-dimensional reconstruction on the image data of the multiple vertebra segments to obtain three-dimensional vertebra images corresponding to the multiple vertebra segments;

a position information recording unit 223, configured to identify and record position information of a key feature point corresponding to each vertebra segment in the scanned image according to vertebra features;

a local coordinate system establishing unit 224, configured to establish a local coordinate system corresponding to the plurality of vertebra segments according to the position information;

a surgical path keypoint locating unit 225 for locating a surgical path keypoint in a local coordinate system corresponding to a plurality of said vertebral segments, based on a plurality of said three-dimensional vertebral images;

a path planning unit 226, configured to plan a surgical path of a corresponding vertebral segment according to the surgical path key points corresponding to a plurality of the vertebral segments.

In one embodiment of the present application, the instance partitioning unit 221 includes:

the example segmentation subunit is used for carrying out vertebra example segmentation processing on the scanning image through the trained neural network model;

the label searching subunit is used for searching a label corresponding to the currently segmented vertebra segment according to the vertebra characteristics;

a label marking subunit, configured to mark the currently segmented vertebra segment with the label;

and the iterative processing subunit is used for performing iterative optimization processing on the marked vertebra segment as the input of the neural network model and outputting the next vertebra segment until image data of a plurality of different vertebra segments in the target vertebra are obtained.

In another embodiment of the present application, the instance splitting unit 221 further includes:

and the first image data extraction subunit is used for extracting the image data of the corresponding vertebra segment from the scanned image according to the label corresponding to the marked vertebra instance.

In another embodiment of the present application, the image data extracting subunit includes:

an image noise removing subunit, configured to remove noise in the scanned image to obtain a first spine image, where the noise is image data in the scanned image except for the vertebral segment;

the image resampling sub-unit is used for resampling the image data in the first spine image to obtain a second spine image in the direction of preset resolution;

and the second image data extraction subunit is used for extracting the image data of the corresponding vertebra segment from the second vertebra image according to the label corresponding to the marked vertebra instance.

In another embodiment of the present application, the image noise removing subunit includes:

the image preprocessing subunit is used for carrying out image preprocessing on the scanned image to obtain a third spine image, wherein the image preprocessing comprises gray level conversion, window adjustment and threshold value division processing;

a connected region area calculation subunit, configured to calculate an area of a connected region existing in the third spine image, so as to obtain an area of at least one connected region;

a vertebra region determining subunit, configured to determine a vertebra region in the third vertebra image according to at least one of the communication area areas;

and the third image data extraction subunit is used for extracting the image data corresponding to the vertebra region from the third vertebra image to obtain the first vertebra image.

In a specific embodiment of the present application, the connected component area calculating subunit is specifically configured to:

traversing the third spine image to calculate the area of the communication area by using a communication area extraction calculation formula, wherein the communication area extraction calculation formula is as follows:

Until B_i+1＝＝B_i

wherein A is a communication region, B_iRepresenting a certain pixel point in the communication area A, wherein i is an integer and is more than or equal to 0; the operators used for the representation of values and operations are in the form of a 3 x 3 squared matrix, filled with the number 1.

In another embodiment of the present application, the connected component area calculating subunit is further specifically configured to:

and determining a vertebra region in the third vertebra image according to a first large communication region, wherein the first large communication region is a communication region corresponding to the largest area in the areas of the communication regions.

The key feature points include a vertebral canal center point and a vertebral body center point.

In one embodiment of the present application, the position information recording unit 223 includes:

the vertebra cross section image intercepting subunit is used for intercepting a vertebra cross section image from the scanning image through a first preset cross section;

the vertebral canal region searching subunit is used for searching a vertebral canal region to be identified from the cross-sectional image of the vertebra according to a second large communication region, wherein the second large communication region is a communication region corresponding to a second large area in the area of the communication region;

the vertebral canal center point calculation subunit is used for performing area integration and center solution on the vertebral canal region to obtain a vertebral canal center point in the vertebral cross section image;

and the vertebral body center point calculation subunit is used for obtaining the vertebral body center point of the vertebral body region in the vertebral cross section image through a maximum inscribed circle geometric calculation method.

In an embodiment of the present application, the local coordinate system establishing unit 224 includes:

an origin position determining subunit, configured to determine spinal canal center point position information serving as an origin of the local coordinate system;

the reference plane determining subunit is used for determining a reference plane of the local coordinate system according to the position information of the central point of the vertebral canal and the position information of the central point of the vertebral body;

and the local coordinate system establishing subunit is used for establishing a local coordinate system corresponding to the vertebra segment according to the origin and the reference plane.

The reference plane includes a first reference plane, a second reference plane, and a third reference plane, and the reference plane determination subunit includes: the vertebra axis determining subunit is used for determining a vertebra axis according to the position information of the central point of the vertebral canal and the position information of the central point of the vertebral body by utilizing a least square fitting linear formula of coordinate points in space;

a first plane determining subunit for taking a plane perpendicular to the vertebra axis as a first reference plane of the local coordinate system;

the second plane fitting subunit is used for fitting a second reference plane meeting a preset condition according to a first parameter of a plane system equation meeting an optimization condition, wherein the preset condition is a distance condition between the axis of the vertebra and the position information of the central point of the vertebra;

a third plane determination subunit for determining a third reference plane of the local coordinate system based on the first reference plane and the second reference plane.

In an embodiment of the present application, the surgical path keypoint locating unit 225 includes:

a first section intercepting subunit, configured to intercept a first section of the three-dimensional vertebra image through a vertebra auxiliary section, where the first section is a plurality of sections obtained by intercepting the vertebra auxiliary section from a corresponding vertebra segment along a predetermined direction;

a region division cross section subunit, configured to divide the first cross section according to a first region and a second region by a region division method to obtain a set of cross sections, where the set of cross sections includes a second cross section and a third cross section;

a central position point determining subunit, configured to determine a group of central position points corresponding to the group of cross sections by using a method of drawing a maximum inscribed circle;

the central position point determining subunit is used for determining a group of central position points of the vertebral pedicle according to the plurality of groups of central position points, wherein the group of central position points of the vertebral pedicle comprises a first central position point of the vertebral pedicle and a second central position point of the vertebral pedicle;

the path intersection point calculation subunit is used for calculating the path intersection point of the pedicle screw implantation operation through a pedicle intersection point calculation formula;

and the operation path key point determining subunit is used for determining the operation path key point according to the group of the central position points of the vertebral pedicle and the path intersection point.

It should be noted that, because the contents of information interaction, execution process, and the like between the modules are based on the same concept as that of the embodiment of the method of the present application, specific functions and technical effects thereof may be specifically referred to a part of the embodiment of the method, and details are not described here.

Fig. 23 is a schematic diagram of a pedicle screw placement surgical path planning device provided in an embodiment of the present application. As shown in fig. 23, the pedicle screw placing operation path planning device 23 of this embodiment includes: a processor 230, a memory 231, and a computer program 232, such as a speech recognition program, stored in the memory 231 and operable on the processor 230. The processor 230 executes the computer program 232 to implement the steps of the above-mentioned each pedicle screw placement surgical path planning method, such as the steps 101-106 shown in fig. 1. Alternatively, the processor 230 executes the computer program 232 to implement the functions of the modules/units in the device embodiments, such as the functions of the units 221-224 shown in fig. 22.

Illustratively, the computer program 232 may be divided into one or more modules/units, which are stored in the memory 231 and executed by the processor 230 to accomplish the present application. One or more of the modules/units may be a series of computer program instruction segments capable of performing specific functions for describing the execution of the computer program 232 in the pedicle screw placement surgical path planning device 23. For example, the computer program 232 may be divided into an example dividing unit 221, a three-dimensional reconstruction unit 222, a position information recording unit 223, a local coordinate system establishing unit 224, a surgical path key point positioning unit 225, and a path planning unit 226, and specific functions of each unit are described in the embodiment corresponding to fig. 1, which are not described herein again.

The pedicle screw placement surgical path planning device may include, but is not limited to, a processor 230, a memory 231. Those skilled in the art will appreciate that fig. 23 is merely an example of the pedicle screw placement surgical path planning device 23, does not constitute a limitation of the pedicle screw placement surgical path planning device 23, and may include more or fewer components than those shown, or some components in combination, or different components, e.g., the pedicle screw placement surgical path planning device may further include an input-output device, a network access device, a bus, etc.

The Processor 230 may be a Central Processing Unit (CPU), other general purpose Processor, a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), an off-the-shelf Programmable Gate Array (FPGA) or other Programmable logic device, discrete Gate or transistor logic, discrete hardware components, etc. A general purpose processor may be a microprocessor or the processor may be any conventional processor or the like.

The memory 231 may be an internal storage unit of the pedicle screw placement operation path planning device 23, such as a hard disk or a memory of the pedicle screw placement operation path planning device 23. The memory 231 may also be an external storage device of the pedicle screw inserting operation path planning device 23, such as a plug-in hard disk, a Smart Media Card (SMC), a Secure Digital (SD) Card, a Flash Card (Flash Card), or the like provided on the pedicle screw inserting operation path planning device 23. Further, the memory 231 may also include both an internal memory unit and an external memory device of the pedicle screw placement surgical path planning device 23. The memory 231 is used to store computer programs and other programs and data required for pedicle screw placement into the surgical path planning device. The memory 231 may also be used to temporarily store data that has been output or is to be output.

The embodiment of the application also provides a computer readable storage medium, which stores a computer program, and when the computer program is executed by a processor, the method for planning the surgical path for implanting the pedicle screw can be realized.

The embodiment of the application provides a computer program product, and when the computer program product runs on a pedicle screw implantation operation path planning device, the pedicle screw implantation operation path planning method can be realized when the pedicle screw implantation operation path planning device runs.

It will be apparent to those skilled in the art that, for convenience and brevity of description, only the above-mentioned division of the functional units and modules is illustrated, and in practical applications, the above-mentioned function distribution may be performed by different functional units and modules according to needs, that is, the internal structure of the apparatus is divided into different functional units or modules, so as to perform all or part of the functions described above. Each functional unit and module in the embodiments may be integrated in one processing unit, or each unit may exist alone physically, or two or more units are integrated in one unit, and the integrated unit may be implemented in a form of hardware, or in a form of software functional unit. In addition, specific names of the functional units and modules are only for convenience of distinguishing from each other, and are not used for limiting the protection scope of the present application. The specific working processes of the units and modules in the system may refer to the corresponding processes in the foregoing method embodiments, and are not described herein again.

In the above embodiments, the descriptions of the respective embodiments have respective emphasis, and reference may be made to the related descriptions of other embodiments for parts that are not described or illustrated in a certain embodiment.

Those of ordinary skill in the art will appreciate that the various illustrative elements and algorithm steps described in connection with the embodiments disclosed herein may be implemented as electronic hardware or combinations of computer software and electronic hardware. Whether such functionality is implemented as hardware or software depends upon the particular application and design constraints imposed on the implementation. Skilled artisans may implement the described functionality in varying ways for each particular application, but such implementation decisions should not be interpreted as causing a departure from the scope of the present application.

The above-mentioned embodiments are only used for illustrating the technical solutions of the present application, and not for limiting the same; although the present application has been described in detail with reference to the foregoing embodiments, it should be understood by those of ordinary skill in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some technical features may be equivalently replaced; such modifications and substitutions do not substantially depart from the spirit and scope of the embodiments of the present application and are intended to be included within the scope of the present application.

Claims (14)

1. A pedicle screw placement surgical path planning method, the method comprising:

carrying out vertebra example segmentation processing on a scanned image of a target vertebra to obtain image data of a plurality of different vertebra segments in the target vertebra;

performing three-dimensional reconstruction on the image data of the multiple vertebral segments to obtain three-dimensional vertebral images corresponding to the multiple vertebral segments;

according to the vertebra features, identifying and recording position information of key feature points corresponding to each vertebra segment in the scanned image;

establishing a local coordinate system corresponding to a plurality of the vertebra segments according to the position information;

based on a plurality of the three-dimensional vertebral images, locating a surgical path keypoint in a local coordinate system corresponding to a plurality of the vertebral segments;

and planning the surgical path of the corresponding vertebral segment according to the surgical path key points corresponding to the plurality of vertebral segments.

2. The method for planning a surgical path for pedicle screw placement according to claim 1, wherein the performing a vertebral instance segmentation process on the scanned image of the target vertebra to obtain image data of a plurality of different vertebral segments in the target vertebra comprises:

carrying out vertebra example segmentation processing on the scanned image through a trained neural network model;

according to the vertebra characteristics, searching a label corresponding to the vertebra segment obtained by current segmentation;

marking the currently segmented vertebral segment with the label;

and performing iterative optimization processing by taking the marked vertebra segment as the input of the neural network model, and outputting the next vertebra segment until image data of a plurality of different vertebra segments in the target vertebra are obtained.

3. The pedicle screw placement procedure path planning method according to claim 2, wherein the vertebral instance segmentation processing is performed on the scanned image of the target vertebra to obtain image data of a plurality of different vertebral segments in the target vertebra, and comprises:

and extracting image data of corresponding vertebra segments from the scanned image according to the labels corresponding to the marked vertebra instances.

4. A pedicle screw placement surgical path planning method as claimed in claim 3, wherein said extracting image data of corresponding vertebral segments from said scanned image according to the labeled labels corresponding to said vertebral instances comprises:

removing noise in the scanned image to obtain a first vertebra image, wherein the noise is image data except the vertebra segment in the scanned image;

resampling image data in the first spine image to obtain a second spine image in a preset resolution direction;

and extracting image data of corresponding vertebra segments from the second vertebra image according to the labels corresponding to the marked vertebra instances.

5. The pedicle screw placement surgical path planning method according to claim 4, wherein the removing noise from the scan image to obtain a first spinal image comprises:

carrying out image preprocessing on the scanned image to obtain a third spine image, wherein the image preprocessing comprises gray level conversion, window adjustment and threshold division processing;

calculating the area of a connected region existing in the third vertebra image to obtain the area of at least one connected region;

determining a vertebral region in the third vertebral image according to at least one of the communication zone areas;

and extracting image data corresponding to the vertebra region from the third vertebra image to obtain the first vertebra image.

6. The pedicle screw placement procedure path planning method according to claim 5, wherein said calculating the area of the connected region existing in the third vertebral image to obtain at least one area of the connected region comprises:

traversing the third spine image to calculate the area of the communication area by using a communication area extraction calculation formula, wherein the communication area extraction calculation formula is as follows:

Until B_i+1＝＝B_i

wherein A is a communication region, B_iRepresenting a certain pixel point in the communication area A, wherein i is an integer and is more than or equal to 0; the operators used for the representation of values and operations are in the form of a 3 x 3 squared matrix, filled with the number 1.

7. A pedicle screw placement surgical path planning method as claimed in claim 5, wherein said determining a vertebral region in said third spinal image based on at least one of said communication zone areas comprises:

and determining a vertebra region in the third vertebra image according to a first large communication region, wherein the first large communication region is a communication region corresponding to the largest area in the areas of the communication regions.

8. The pedicle screw placement surgical path planning method according to claim 5, wherein the key feature points include a vertebral canal center point and a vertebral body center point; the identifying and recording the position information of the key feature point corresponding to each vertebral level in the scanned image according to the vertebral features comprises the following steps:

intercepting a vertebra cross section image from the scanning image through a preset cross section;

searching a vertebral canal region to be identified from the cross-sectional image of the vertebra according to a second large communication area, wherein the second large communication area is a communication area corresponding to a second large area in the area of the communication area;

performing area integration and center solution on the vertebral canal region to obtain a vertebral canal center point in the vertebral cross section image;

and obtaining the central point of the vertebral body region in the cross section image of the vertebra by a maximum inscribed circle geometric calculation method.

9. The method for planning a surgical path for pedicle screw implantation according to claim 1, wherein the position information includes information on a central point of a vertebral canal and information on a central point of a vertebral body, and the establishing a plurality of corresponding local coordinate systems of the vertebral segments according to the position information includes:

determining vertebral canal center point position information as an origin of the local coordinate system;

determining a reference plane of the local coordinate system according to the position information of the central point of the vertebral canal and the position information of the central point of the vertebral body;

and establishing a local coordinate system corresponding to the vertebra segments according to the origin and the reference plane.

10. A pedicle screw placement procedure path planning method as claimed in claim 9, wherein the reference planes comprise a first reference plane, a second reference plane and a third reference plane, and wherein determining the reference plane of the local coordinate system based on the spinal canal center point location information and the vertebral body center point location information comprises:

determining the axis of the vertebra according to the position information of the central point of the vertebral canal by utilizing a least square fitting linear formula of coordinate points in space;

taking a plane perpendicular to the vertebral axis as a first reference plane of the local coordinate system;

fitting a second reference plane meeting a preset condition according to a first parameter of a plane system equation meeting an optimization condition, wherein the preset condition is a distance condition between the axis of the vertebra and the position information of the central point of the vertebra;

determining a third reference plane of the local coordinate system from the first reference plane and the second reference plane.

11. The pedicle screw placement surgical path planning method according to claim 1, wherein said locating a surgical path keypoint in a local coordinate system corresponding to a plurality of said vertebral segments based on a plurality of said three-dimensional vertebral images comprises:

a first section of the three-dimensional vertebra image is cut through a vertebra auxiliary section, and the first section is a plurality of sections obtained by cutting the vertebra auxiliary section from corresponding vertebra segments along a preset direction;

dividing the first section according to a first region and a second region by a region dividing method to obtain a group of sections, wherein the group of sections comprises a second section and a third section;

determining a group of central position points corresponding to the group of sections by adopting a method for drawing a maximum inscribed circle;

determining a group of pedicle center position points according to the plurality of groups of center position points, wherein the group of pedicle center position points comprise a first pedicle center position point and a second pedicle center position point;

calculating the path intersection point of the pedicle screw implantation operation through a pedicle intersection point calculation formula;

and determining a key point of the operation path according to a group of the central position points of the vertebral pedicle and the path intersection point.

12. A pedicle screw placement surgical path planning device, comprising:

the example segmentation unit is used for carrying out vertebra example segmentation processing on the scanned image of the target vertebra to obtain image data of a plurality of different vertebra segments in the target vertebra;

the three-dimensional reconstruction unit is used for performing three-dimensional reconstruction on the image data of the multiple vertebral segments to obtain three-dimensional vertebral images corresponding to the multiple vertebral segments;

the position information recording unit is used for identifying and recording the position information of the key feature point corresponding to each vertebra segment in the scanned image according to the vertebra features;

the local coordinate system establishing unit is used for establishing a local coordinate system corresponding to the plurality of vertebra segments according to the position information;

a surgical path keypoint locating unit for locating a surgical path keypoint in a local coordinate system corresponding to a plurality of said vertebral segments based on a plurality of said three-dimensional vertebral images;

and the path planning unit is used for planning the operation path of the corresponding vertebral segment according to the operation path key points corresponding to the plurality of vertebral segments.

13. A pedicle screw placement surgical path planning device comprising a processor, a memory and a computer program stored in the memory and executable on the processor, characterized in that the processor, when executing the computer program, implements a pedicle screw placement surgical path planning method as claimed in any one of claims 1 to 11.

14. A computer-readable storage medium storing a computer program, wherein the computer program, when executed by a processor, implements a pedicle screw placement surgical path planning method as claimed in any one of claims 1 to 11.

Images