CN115272615A - Medical semantic segmentation method for femoral point cloud model and application of medical semantic segmentation method - Google Patents
Medical semantic segmentation method for femoral point cloud model and application of medical semantic segmentation method Download PDFInfo
- Publication number
- CN115272615A CN115272615A CN202210953666.3A CN202210953666A CN115272615A CN 115272615 A CN115272615 A CN 115272615A CN 202210953666 A CN202210953666 A CN 202210953666A CN 115272615 A CN115272615 A CN 115272615A
- Authority
- CN
- China
- Prior art keywords
- point cloud
- femoral
- point
- cloud set
- shaft
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T17/00—Three dimensional [3D] modelling, e.g. data description of 3D objects
- G06T17/20—Finite element generation, e.g. wire-frame surface description, tesselation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/11—Region-based segmentation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10028—Range image; Depth image; 3D point clouds
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20021—Dividing image into blocks, subimages or windows
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30004—Biomedical image processing
- G06T2207/30008—Bone
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Computer Graphics (AREA)
- Geometry (AREA)
- Software Systems (AREA)
- Apparatus For Radiation Diagnosis (AREA)
Abstract
The invention discloses a medical semantic segmentation method for a femoral point cloud model, which comprises the following steps: the method comprises the following steps: for the collected femur model point cloud collectionPCarry out femur dry point cloud collectionP shaft Extracting; step two: for relabeled point cloud setP upper Extracting the Point cloud of the femoral headP head (ii) a Step three: extracting femoral trochanter point cloud setP rotor And femoral neck point cloud collectionP neck (ii) a Step four: extracting point cloud set of medial condyle of femurP inter condylar And lateral condyle area point cloud setP lateral condylar . Meanwhile, the invention discloses application of the medical semantic segmentation method of the femoral point cloud model in femoral parametric measurement, personalized implant design and operation planning. The invention realizes the medical semantic segmentation of the femoral point cloud model, and has the advantages of simplicity, high efficiency andthe method has the characteristics of practicability and the like, and has important significance for the digital diagnosis and treatment of the orthopedics department.
Description
Technical Field
The invention relates to a medical semantic segmentation method of a femoral point cloud model and application thereof, belonging to the technical field of digital orthopedics.
Background
The femur is the longest bone of the human body and is divided into two ends. The proximal end of the femur is circular, the femoral head and the acetabulum form a hip joint, the distal end of the hip joint is expanded to the left and the right to form a lateral condyle and a medial condyle, and the lateral condyle and the medial condyle are connected with the patella. The femoral injuries comprise fracture of femoral shaft, femoral head necrosis, internal and external condylar and other regional damages, are common diseases and frequently encountered diseases in clinical orthopedics, and particularly have the highest incidence rate in middle-aged and old patients. The medical semantic segmentation of the femur model is important basic work of orthopedic diagnosis and treatment such as femur parameter measurement, personalized implant design and operation planning.
The point cloud model area segmentation is to segment a point cloud set into a plurality of mutually disjoint subsets, and the point clouds with the same characteristics form a subset corresponding to an area. At present, a point cloud segmentation method is mainly based on differential geometric information such as normal vectors, gaussian curvatures, mean curvatures and the like, and divides a point set with the same differential geometric attributes into a subset. The differential geometric information such as normal vector, gaussian curvature and average curvature needs secondary calculation, the calculation efficiency is low, and calculation errors exist. More importantly, the existing point cloud segmentation method only considers differential geometric information of a point cloud set, fails to consider medical anatomical features of a femur model, cannot realize point cloud segmentation with medical semantic information, and cannot support the work of femur parametric measurement, personalized implant design, surgical planning and the like. Therefore, the invention provides an efficient point cloud segmentation method, which is used for realizing medical semantic segmentation of a femoral point cloud model.
Disclosure of Invention
The invention aims to solve the technical problem of providing a medical semantic segmentation method of a femoral point cloud model, which provides scientific basis for femoral parametric measurement, personalized implant design, operation planning and the like and has important significance for digital orthopedic diagnosis and treatment.
Meanwhile, the invention provides application of the medical semantic segmentation method of the femoral point cloud model.
In order to solve the technical problems, the technical scheme adopted by the invention is as follows:
a medical semantic segmentation method for a femoral point cloud model comprises the following steps:
the method comprises the following steps: carrying out femur trunk point cloud set P on the collected femur model point cloud set P shaft Obtaining a femur dry point cloud set P shaft Relabeled point cloud set P upper And a relabeled point cloud set P lower (ii) a Relabeled point cloud set P upper Comprises a femoral head, a femoral neck and a femoral trochanter; relabeled point cloud set P lower Comprises a medial femoral condyle and a lateral femoral condyle;
step two: for relabeled point cloud set P upper Extracting femoral head Point cloud set P head ；
Step three: extracting femur rotor point cloud set P rotor And femur neck point cloud set P neck ；
Step four: extracting point cloud set P of medial condyle of femur inter condylar And lateral condyle area point cloud set P lateral condylar 。
The first step comprises the following steps:
step 1a: segmenting the femur model point cloud set P into an initial point cloud set P upper And P lower ；
Step 1b: extracting point cloud set P upper Point cloud collection P of upper segment of middle femoral shaft upper-shaft ；
Step 1c: extracting point cloud set P lower Middle femoral shaft lower segment point cloud set P lower-shaft ；
Step 1d: merging point cloud set P upper-shaft And point cloud set P lower-shaft To complete the femur trunk point cloud set P shaft And (4) extracting.
In step 1a, an initial set of point clouds P upper And P lower The segmentation method comprises the following steps: constructing a femur point cloud model axial alignment bounding box AABB, and setting the axial direction of a femur model point cloud set P as a Y coordinate axis; counting the maximum value Ymax and the minimum value Ymin of the femur model point cloud set P in the Y coordinate axis direction, and centering the point in the Y coordinate axis direction
Dividing the femur point cloud model into two point cloud sets P upper And P lower ；
Wherein an initial set of point clouds P upper Comprises a femoral shaft upper section, a femoral head, a femoral neck and a femoral trochanter, and an initial point cloud set P lower Comprises a lower femoral shaft segment, a medial femoral condyle and a lateral femoral condyle.
In step 1b, an initial point cloud set P is extracted upper The method for collecting the femur trunk point cloud comprises the following steps: according to the characterization index CIS of the femoral stem, an initial point cloud set P is obtained upper Performing binary iteration segmentation;
the method comprises the following steps of: firstly, a point cloud slice with the thickness of delta at the Y coordinate axis Y = Y is obtained, namely [ Y-delta/2, Y + delta/2]An interior point cloud slice, then projecting the point cloud slice to plane Y = Y; then, calculating the centroid of the projection point cloud of the slice
Wherein P is i ＝(x i ,y i ,z i )，i＝1,2,…,n， P i Point coordinates, x, for point cloud slices i ,y i ,z i Are respectively a point P i The coordinate value of XYZ axis of (i) is the number of points; finally, point P is calculated i From the point cloud centroid P c Maximum value of distance D max And a minimum value D min Then CIS = D max -D min ；
The binary iterative segmentation is based on the Y-axis direction of the femur point cloud model
Whether CIS at the position (i =1,2,3, \8230;, n) is larger than a specified threshold value or not is judged, and a femoral shaft point cloud subset, namely a femoral shaft upper segment point cloud set P is extracted upper-shaft I is the number of binary iterative segmentations, yi is the initial point cloud set P upper Is divided into two halves between the maximum value Ymax and the minimum value Ymin;
wherein, the threshold value range: 5-8mm;
repeatedly executing binary iterative segmentation until the binary iterative segmentation point is smaller than the point cloud slice thickness delta and the upper segment point cloud collection P of the femur stem upper-shaft Is extracted, at which point a set of relabeled points P upper The middle part comprises the femoral head, the femoral neck and the femoral trochanter;
in step 1c, an initial point cloud set P is collected lower Rotating pi around the Z axis, executing the step 1b again, and extracting the cloud set P of the lower point of the femoral shaft lower-shaft (ii) a At this point, the relabeled point cloud set P lower Including the medial and lateral femoral condyles.
The second step comprises:
step 2a: relabeling-based point cloud set P upper Constructing a point cloud fitting sphere, and extracting a femoral head initial point cloud set P in ；
And step 2b: extracting the initial point of the femoral head in And judging the type of the point cloud boundary;
and step 2c: executing breadth-first k adjacent region growing algorithm to extract femoral head point cloud set P head 。
In step 2a, in the relabeled point cloud set P upper In (1), selecting a non-coplanar point P 0 、P 1 、P 2 And P 3 Construction of the center of sphere as (x) 0 ,y 0 ,z 0 ) An initial fitting sphere S with radius R, whose equation is (x-x) 0 ) 2 +(y-y 0 ) 2 +(z-z 0 ) 2 ＝R 2 (ii) a Then, a re-labeled point cloud set P is calculated upper Point p (x) of i ,y i ,z i ) Distance to the sphere S
If d is<Dividing the threshold value, marking the point as a spherical fitting point P if the value of the division threshold value ranges from 1 mm to 2mm a ；
Repeating the above steps n times, wherein n is not less than 1800, then the point cloud set P of the relabel is obtained upper Point cloud set P with upper femoral head point set being spherically fitted and re-marked upper The fitting spherical surface point cloud set is marked as the initial point cloud set P of the femoral head in The rest point cloud sets are marked as P out ；
In step 2b, the point cloud boundary is divided into an adjacent boundary AD and an internal boundary ID;
the abutment boundary AD is the boundary between the femoral head and the femoral neck;
the internal boundary ID is the femoral head point cloud not extracted by step 2a, resulting from the local concavity of the femoral head surface;
adjacent boundary AD belongs to point cloud set P in And with P out The extraction basis of the adjacent point sets is as follows: order to
N k (P a ) Is a reaction with P a K is 15 if the point set is formed by k points nearest to the point set
Then P is a Set P for point cloud in The boundary points of (a); the boundary point set is an adjacent boundary AD;
the method for judging the internal boundary ID comprises the following steps: as can be seen from the femoral head convex-concave region, if the number of boundary points of the convex-concave region is less than 18, the boundary is the inner boundary ID.
In step 2c, the breadth-first k-nearest neighbor region growing algorithm is as follows:
from point P e Starting from the point of view, extracting k neighbor point sets N k (P e ) And adding itCloud set P added to initial point of femoral head in Then recursively extracts N k (P e ) K neighbor set N of the most distant point P' from the center in the intersection of (2) k (P') up to
Taking the internal boundary ID in the step 2b as a clue, repeatedly executing breadth-first k neighbor region growing, extracting the femoral head point cloud inside the internal boundary ID and not extracted in the step 2a, and supplementing the femoral head point cloud to the femoral head initial point cloud set P in And completing the femoral head point cloud set P until the femoral head point cloud which is not extracted does not exist any more head And (6) dividing.
The third step comprises:
step 3a: extracting femoral trochanter region characteristic line cluster P clue ；
And step 3b: extracting a femoral trochanter point set P by taking the characteristic line as a clue rotor 。
Characteristic line cluster P clue The device consists of a horizontal characteristic line and a vertical characteristic line;
the extraction process of the horizontal characteristic line and the vertical characteristic line comprises the following steps: firstly, making a point cloud slice which is vertical to an X axis and has the thickness delta; then, extracting a central axis of the point cloud slice; finally, according to the femoral trochanteric characteristics, taking the maximum value point in the Y direction of the central axis of the point cloud slicing as a segmentation point, and obtaining a current position characteristic line;
taking the current position characteristic line as a clue, executing the breadth-first k adjacent region growing algorithm in the step 2c to complete the extraction of the femoral trochanter point cloud set; obtaining femoral trochanter point cloud set P rotor And femur neck point cloud set P neck 。
The method for obtaining the segmentation characteristic points of the medial femoral condyle and the lateral femoral condyle comprises the following steps of: firstly, carrying out point cloud slicing in the vertical direction; then, solving a midpoint value mid in the X-axis direction of a point cloud slice characteristic line, wherein the point cloud slice characteristic line is the central axis of the point cloud slice; finally, the minimum value of the inner Y axis of the interval mid-5 and x not more than mid +5 is obtained, namely the inner and outer condyle segmentation characteristic point P b ；
According to the segmentation feature point P b1 (x 1 ,y 1 ,z 1 )、P b2 (x 2 ,y 2 ,z 2 )、P b3 (x 3 ,y 3 ,z 3 ) Determining a plane Π: ax + By + Cz + D =0, computing a set P of relabeled point clouds lower Point P of b (x i ,y i ,z i ) Pi distance to plane
Wherein A, B, C and D are plane equation coefficients; setting a threshold value to be approximately equal to 0, and extracting a point cloud set P marked again lower The inner and outer condyles are divided into cloud sets of point to form dividing characteristic lines;
using the segmented feature lines to re-label the point cloud set P lower Segmented into femoral medial condyle point cloud set P inter condylar And a lateral condyle area point cloud set P lateral condylar 。
An application of a medical semantic segmentation method of a femoral point cloud model in femoral parametric measurement, personalized implant design and operation planning.
To this end, the femoral point cloud model is segmented into six subsets of femoral shaft, femoral head, femoral neck, femoral trochanter, medial condyle and lateral condyle.
The dichotomous iterative segmentation defines a femoral Shaft Characterization Index CIS (Characterization Index of Shaft) according to the medical anatomical features of the femur, which characterizes the geometrical shape of the cross section of the point cloud, and realizes a femoral Shaft point cloud set P by using the CIS of the dichotomous iterative segmentation and the segmentation position shaft And (4) quickly extracting.
The invention judges and extracts the point cloud set P by using whether the k neighbor intersection is an empty set or not in A middle boundary point; at the same time, it is divided into a contiguous boundary and an inner boundary according to its formation reason, wherein the inner boundary is a clue for extracting a set of femoral head points that have not been extracted yet.
The method starts from the medical anatomical features of the femur, adopts methods such as point cloud slicing, slice projection, axis extraction, feature point extraction and k-nearest neighbor region growth, and not only avoids the problems of large calculated amount and calculation error caused by the dependence of the existing method on differential geometric information such as normal vector, curvature and the like, but also realizes the medical semantic segmentation of the femur model.
The beneficial effects of the invention are: the invention provides a thighbone point cloud model segmentation method combined with prior medical knowledge. The method has the advantages of simplicity, high efficiency, practicality and the like, can provide scientific basis for parametric measurement, implant customized design and operation planning, and has important significance for the implementation of digital diagnosis and treatment of orthopedics.
Drawings
FIG. 1 is a schematic diagram of the present invention;
FIG. 2 is a femoral point cloud collection diagram;
FIG. 3 is a representation CIS chart of the femoral shaft characterization index;
FIG. 4 is a bipartite iterative slice segmentation graph;
FIG. 5 is a point cloud set P of the upper part of the femoral shaft upper-shaft A drawing;
FIG. 6 is a cloud set P of points on the lower part of the femoral shaft lower-shaft A drawing;
FIG. 7 is a femoral stem point cloud set P shaft Drawing;
FIG. 8 is a cloud set P of femoral head initial points in A drawing;
FIG. 9 is a point cloud boundary map;
FIG. 10 is a schematic of a boundary point;
FIG. 11 is a femoral head point cloud set P head Drawing;
FIG. 12 is a femoral trochanter characteristic cluster diagram;
FIG. 13 is a schematic view of femoral trochanter feature line extraction;
FIG. 14 is a femoral trochanter point cloud set P rotor And femur neck point cloud set P neck Drawing;
FIG. 15 is a schematic view of segmentation feature point extraction for the medial and lateral condyles;
FIG. 16 is a medial-lateral condyle segmentation feature point diagram and a medial-lateral condyle segmentation line diagram;
FIG. 17 is a point cloud P of medial femoral condyle inter condylar And lateral condyle point cloudSet P lateral condylar Drawing.
Detailed Description
The present invention will be further described with reference to the accompanying drawings.
As shown in fig. 1, a medical semantic segmentation method for a femoral point cloud model includes the following steps:
the method comprises the following steps: extracting femur dry point cloud set P shaft The method comprises the following specific steps:
step 1a: dividing the femur model point cloud set P into point cloud sets P upper And P lower ；
As shown in fig. 2, a femur model point cloud model axial alignment bounding box AABB is constructed, and the femur model point cloud set P axial direction is set as a Y coordinate axis; counting the maximum value Ymax and the minimum value Ymin of the femur model point cloud set in the Y coordinate axis direction, and midpoint in the Y direction
Dividing the femur point cloud model into two point cloud sets P upper And P lower . The maximum value Ymax and the minimum value Ymin are the maximum value and the minimum value of the femoral point cloud model in the axial direction.
Wherein a point cloud P upper Comprises the areas of the upper part of the femoral shaft, the femoral head, the femoral neck, the femoral trochanter and the like, and a point cloud set P lower Comprises a lower femoral shaft segment, a medial femoral condyle and a lateral femoral condyle. Then, respectively to P upper And P lower Executing the steps 1b to 1d, and extracting a femur trunk point cloud set P shaft 。
Step 1b: extracting point cloud set P upper A middle femur stem point cloud set;
according to a stock skeleton Characterization Index CIS (characteristics Index of Shaft), a point cloud set P is determined upper And (5) performing binary iteration segmentation. Wherein, the CIS description point cloud set P is represented by a femoral shaft representation index upper Axial cross-sectional shape. As shown in fig. 3, the CIS solving steps are: firstly, a point cloud slice of the thickness delta at the Y axis Y = Y is obtained, namely [ Y-delta/2, Y + delta/2]An interior point cloud slice, then projecting the point cloud slice to plane Y = Y; then, calculating the center of mass of the slice projection point cloud
Wherein P is i ＝(x i ,y i ,z i )，i＝1,2,…,n，P i Point coordinates, x, for point cloud slices i ,y i ,z i Are respectively a point P i The coordinate value of XYZ axis of (i) is the number of points; finally, calculating the cloud distance centroid P of the points c Maximum value of distance D max And a minimum value D min Then CIS = D max -D min 。
The binary iterative segmentation is based on the Y-axis direction of the femur point cloud model
Whether the CIS is larger than a specified threshold value at the position (i =1,2,3, \ 8230;, n) or not is judged, a femoral shaft point cloud subset is extracted, i is the number of times of two-iteration segmentation, yi is the halving segmentation between a maximum value Ymax and a minimum value Ymin, and is a model coordinate, wherein the maximum value Ymax and the minimum value Ymin are P in FIG. 2 upper Maximum and minimum values of (a); wherein, the threshold is set according to the cross section shape of the femoral shaft, and the value range is as follows: 5-8mm. As shown in fig. 4, in fig. (a), when i =1, P is upper Is less than a specified threshold, then P upper Lower segment is femur dry point cloud set P upper-shaft ，P upper The upper point cloud set is relabeled as P upper (ii) a In diagram (b), P upper I =3, CIS is less than the specified threshold, when P is present upper The lower segment is a femoral shaft point cloud subset, P upper The upper point cloud subset is relabeled as P upper (ii) a And (c) extracting the femoral shaft point cloud subset by the third binary iteration segmentation.
Repeatedly executing binary iteration segmentation until binary iteration cutting point and P upper Is overlapped, i.e. less than the point cloud slice thickness delta. As shown in FIG. 5, a point cloud P is obtained from the upper portion of the femoral shaft upper-shaft Is extracted, at this time, P upper Including the femoral crown, femoral neck, and femoral trochanter regions.
Step 1c: extracting point cloud set P lower A middle femur stem point cloud set;
as shown in FIG. 6, a point cloud P is collected lower Rotating pi around the Z axis, executing the step 1b again, and extracting the lower segment point cloud set P of the femoral stem lower-shaft 。
Specifically, a point cloud set P is extracted lower A middle femur trunk point cloud set;
rotating point cloud set P according to the stock skeleton Characterization Index CIS (Characterization Index of Shaft) lower And (5) performing binary iterative segmentation.
Binary iterative segmentation, i =1, P lower Is less than a specified threshold, then P lower The lower segment is a femur dry point cloud set P lower-shaft ，P lower The upper point cloud set is relabeled as P lower (ii) a Relabeled P lower I =3, CIS is less than the specified threshold, when P is present lower The next segment is a femoral shaft point cloud subset, P lower The upper point cloud subset is relabeled as P lower (ii) a For P re-marked again lower And performing third binary iterative segmentation on the extracted femoral shaft point cloud subset.
Repeatedly executing binary iteration segmentation until binary iteration cutting point and P lower Is overlapped, i.e. less than the point cloud slice thickness delta. Lower point cloud set P of femoral shaft lower-shaft Is extracted, at this time, P lower Including a point cloud P of medial femoral condyle inter condylar And lateral condylar point cloud set P lateral condylar A region.
Step 1d: will P upper-shaft And P lower-shaf Merging point cloud sets to complete femur trunk point cloud set P shaft Extracting;
as shown in FIG. 7, P is upper-shaft And P lower-shaf Merging point clouds in the stock skeleton region of the point cloud collection into P shaft And finishing the segmentation of the femur trunk point cloud set. Relabeling the remaining point clouds as P upper And P lower As a point cloud set for subsequent steps.
Step two: extracting femoral head Point cloud set P head The method comprises the following specific steps：
Step 2a: constructing a point cloud fitting sphere, and extracting a femoral head initial point cloud set P in 。
Point cloud set P relabeled at last upper In (1), non-coplanar point P is randomly selected 0 、P 1 、P 2 、P 3 Construction of the center of sphere as (x) 0 ,y 0 ,z 0 ) An initial fitting sphere S with radius R, the equation of which is (x-x) 0 ) 2 +(y-y 0 ) 2 +(z-z 0 ) 2 ＝R 2 (ii) a Then, a point cloud set P is calculated upper Point p (x) of i ,y i ,z i ) Distance to the sphere S
If d is<A segmentation threshold value, the value of which ranges from 1 mm to 2mm according to the geometrical shape of the femoral head, then the point is marked with a spherical fitting point P a 。
Repeating the steps for n times (experiments show that the value of n is 1800, the best fitting effect can be achieved), and then obtaining the point cloud set P upper The upper femoral head point set is sphere fitted. As shown in FIG. 8, the final relabeled point cloud set P upper The fitting spherical point cloud set is marked as the femoral head initial point cloud set P in The rest point cloud sets are marked as P out 。
That is, in step 2a, according to the prior knowledge of femur medical science, an iterative four-point fitting spherical method is adopted, and when the number of spherical fitting points is greater than or equal to a threshold value, the femoral head initial point cloud set P is completed in And (4) extracting.
And step 2b: extracting the initial point cloud set P of the femoral head in And judging the type of the point cloud boundary;
as shown in fig. 9, the point cloud boundary is divided into an adjoining boundary AD (adjacentborder) and an internal boundary ID (Inner Border). The abutment boundary AD is the boundary between the femoral head and the femoral neck; the internal boundary ID represents the femoral head point cloud not extracted by step 2a, which is caused by the local convexity of the femoral head surface. In order to segment a femoral head point cloud set, a point cloud boundary needs to be identified first, and then an adjacent boundary and an inner boundary are distinguished.
As shown in FIG. 10, the point cloud boundary belongs to a point cloud set P in And with P out The extraction basis of the adjacent point sets is as follows: order to
Is a reaction with P a K is 15 if the point set is formed by k points nearest to the point set
Then P is a As a set of points P in The boundary point of (2). And the boundary point set is the point cloud adjacent boundary.
Since the inner boundary ID is caused by local convexo-concave of the femoral head surface, its area is small, so the judgment method is: as can be seen from the femoral head convex-concave area, if the number of boundary points is less than 18, the boundary is an internal boundary.
And step 2c: executing breadth-first k adjacent region growing algorithm to complete femoral head point cloud set P head And (6) dividing.
The idea of breadth-first k-nearest neighbor region growing is as follows:
from point P e Starting from the point of view, extracting k neighbor point sets N k (P e ) And adding P thereto in Then recursively extracts N k (P e ) K nearest neighbor point set N of the farthest point P' from the center in the intersection of (2) k (P') up to
P e Is the point in the internal boundary ID in step 2 b.
Repeatedly performing breadth-first K neighbor region growing with the internal boundary ID as a clue, extracting the femoral head point cloud which is not extracted in the step 2a and is inside the internal boundary ID, and supplementing the femoral head point cloud to P in Point cloud collection until there is no more femoral head point cloud which is not extracted, and then the femoral head point cloud collection P is completed head And (6) dividing. As shown in FIG. 11, the above-mentioned operations are performed to obtain a point cloud P head Is collected from the point cloud P upper And (6) dividing.
Step three: extracting femoral trochanter point cloud set P rotor And femur neck point cloud set P neck The method comprises the following specific steps:
step 3a: extracting femoral trochanter region characteristic line cluster P clue ；
As shown in fig. 12, the cluster of features is composed of horizontal and vertical features, implying the overall boundary and shape profile of the rotor region.
As shown in fig. 13, the extraction process of the feature line is: firstly, making point cloud slices which are perpendicular to an X axis and have the thickness of delta; then, extracting the central axis of the slice; and finally, according to the femoral trochanteric characteristics, taking the maximum value point of the slicing medial axis in the Y direction as a segmentation point, and obtaining a current position characteristic line.
In step 3a, the characteristic line cluster is a point cloud set P marked again according to the rotor region anatomical characteristics upper The horizontal and vertical characteristic lines of (a), which implies the overall boundary and shape profile of the rotor region.
And step 3b: extracting femoral trochanter point cloud set P by taking characteristic line as clue rotor 。
And (4) taking the characteristic line as a clue, and executing the breadth-first k neighbor region growing algorithm in the step 2c again to finish the extraction of the femoral trochanter point cloud set. As shown in FIG. 14, a cloud of points P upper Is divided into a rotor point cloud set P rotor And femur neck point cloud set P neck 。
Will go through the above steps to collect a point cloud P upper Is divided into medical anatomical region point cloud sets of femoral head, femoral neck, femoral trochanter, femoral shaft and the like.
Next, with P relabeled last lower And segmenting the point cloud of the inner and outer condylar areas of the data source.
Step four: extracting point cloud set P of medial condyle of femur inter condylar And lateral condyle point cloud set P lateral condylar The method comprises the following specific steps:
step 4a: extracting the segmentation characteristic points of the medial condyle and the lateral condyle;
as shown in fig. 15, the local extreme points of the characteristic points of the medial condyle and the lateral condyle areas are the minimum values of the Y-axis of the characteristic line, and the solving step is as follows: firstly, slicing a point cloud vertical to a vertical point; then, solving a midpoint value mid in the X-axis direction of the slice characteristic line; and finally, solving the minimum value of the inner Y axis of the interval mid-5 and x not more than mid +5, namely the inner and outer condyle segmentation characteristic points.
In step 4a, according to the anatomical features of the medial and lateral condyles, local extreme points are selected as the segmentation feature points of the medial and lateral condyles.
And 4b: constructing a segmentation plane according to the segmentation feature points, and extracting segmentation;
according to the segmentation feature point P b1 (x 1 ,y 1 ,z 1 )、P b2 (x 2 ,y 2 ,z 2 )、P b3 (x 3 ,y 3 ,z 3 ) Determining a plane Π: ax + By + Cz + D =0, calculating a point cloud set P lower Point P of b (x i ,y i ,z i ) Pi distance to plane
Wherein, A, B, C and D are plane equation coefficients. Setting the threshold value to be approximately equal to 0, and extracting P lower The inner and outer condyles of the point cloud set are divided into point cloud sets to form division characteristic lines. As shown in fig. 16, the segmentation characteristic line between the medial and lateral condyles was formed using a plane fit to the segmentation points.
In step 4b, a three-point plane determining method is utilized to construct a segmentation plane pi, and P on the plane pi is extracted according to the distance from the point to the plane pi lower And (4) dot forming a segmentation characteristic line. The segmentation feature lines describe the regional boundaries of the medial and lateral condylar point cloud sets.
And 4c: using the segmentation feature line to segment the point cloud set P lower 。
As shown in fig. 17, a point cloud set P is formed by dividing a feature line lower Divided into medial condyle P inter condylar And lateral condylar point cloud set P lateral condylar Two point cloud sets.
To this end, the point cloud model of the femur is divided into six subsets of the femoral shaft, the femoral head, the femoral neck, the trochanter of the femur, the medial condyle of the femur and the lateral condyle of the femur. The data size of each subset is smaller, and the detailed modeling representation and the real-time dynamic rendering of the femur model are facilitated. The accurate modeling representation can not only improve the parametric measurement precision, such as the area measurement of the femoral head necrosis lesion region, but also provide a more accurate implant binding surface for the personalized implant design; the real-time dynamic rendering can improve the immersion feeling of the operation planning, thereby improving the quality of the operation diagnosis and treatment.
In the description provided herein, numerous specific details are set forth. It is understood, however, that embodiments of the invention may be practiced without these specific details. In some instances, well-known methods, structures and techniques have not been shown in detail in order not to obscure an understanding of this description.
Similarly, it should be appreciated that in the foregoing description of exemplary embodiments of the invention, various features of the invention are sometimes grouped together in a single embodiment, figure, or description thereof for the purpose of streamlining the disclosure and aiding in the understanding of one or more of the various inventive aspects. However, the disclosed method should not be construed to reflect the intent: that the invention as claimed requires more features than are expressly recited in each claim. Rather, as the following claims reflect, inventive aspects lie in less than all features of a single foregoing disclosed embodiment. Thus, the claims following the detailed description are hereby expressly incorporated into this detailed description, with each claim standing on its own as a separate embodiment of this invention.
Those skilled in the art will appreciate that the modules or units or groups of devices in the examples disclosed herein may be arranged in a device as described in this embodiment, or alternatively may be located in one or more different devices than the devices in this example. The modules in the foregoing examples may be combined into one module or may be further divided into a plurality of sub-modules.
Those skilled in the art will appreciate that the modules in the device in the embodiments may be adaptively changed and disposed in one or more devices different from the embodiments. Modules or units or groups in embodiments may be combined into one module or unit or group and may further be divided into sub-modules or sub-units or sub-groups. All of the features disclosed in this specification (including any accompanying claims, abstract and drawings), and all of the processes or elements of any method or apparatus so disclosed, may be combined in any combination, except combinations where at least some of such features and/or processes or elements are mutually exclusive. Each feature disclosed in this specification (including any accompanying claims, abstract and drawings) may be replaced by alternative features serving the same, equivalent or similar purpose, unless expressly stated otherwise.
Moreover, those skilled in the art will appreciate that although some embodiments described herein include some features included in other embodiments, not other features, combinations of features of different embodiments are meant to be within the scope of the invention and form different embodiments. For example, in the following claims, any of the claimed embodiments may be used in any combination.
Furthermore, some of the described embodiments are described herein as a method or combination of method elements that can be performed by a processor of a computer system or by other means of performing the described functions. A processor with the necessary instructions for carrying out the method or the method elements thus forms a device for carrying out the method or the method elements. Further, the elements of the apparatus embodiments described herein are examples of the following apparatus: the means for performing the functions performed by the elements for the purpose of carrying out the invention.
The various techniques described herein may be implemented in connection with hardware or software or, alternatively, with a combination of both. Thus, the methods and apparatus of the present invention, or certain aspects or portions thereof, may take the form of program code (i.e., instructions) embodied in tangible media, such as floppy diskettes, CD-ROMs, hard drives, or any other machine-readable storage medium, wherein, when the program is loaded into and executed by a machine, such as a computer, the machine becomes an apparatus for practicing the invention.
In the case of program code execution on programmable computers, the computing device will generally include a processor, a storage medium readable by the processor (including volatile and non-volatile memory and/or storage elements), at least one input device, and at least one output device. Wherein the memory is configured to store program code; the processor is configured to perform the method of the invention according to instructions in said program code stored in the memory.
By way of example, and not limitation, computer readable media may comprise computer storage media and communication media. Computer-readable media includes both computer storage media and communication media. Computer storage media store information such as computer readable instructions, data structures, program modules or other data. Communication media typically embodies computer readable instructions, data structures, program modules or other data in a modulated data signal such as a carrier wave or other transport mechanism and includes any information delivery media. Combinations of any of the above are also included within the scope of computer readable media.
As used herein, unless otherwise specified the use of the ordinal adjectives "first", "second", "third", etc., to describe a common object, merely indicate that different instances of like objects are being referred to, and are not intended to imply that the objects so described must be in a given sequence, either temporally, spatially, in ranking, or in any other manner.
While the invention has been described with respect to a limited number of embodiments, those skilled in the art, having benefit of this disclosure, will appreciate that other embodiments can be devised which do not depart from the scope of the invention as disclosed herein. Moreover, it should be noted that the language used in the specification has been principally selected for readability and instructional purposes, and may not have been selected to delineate or circumscribe the inventive subject matter. Accordingly, many modifications and variations will be apparent to those of ordinary skill in the art without departing from the scope and spirit of the appended claims. The present invention has been disclosed with respect to the scope of the invention, which is to be considered as illustrative and not restrictive, and the scope of the invention is defined by the appended claims.
Claims (10)
1. A medical semantic segmentation method of a femoral point cloud model is characterized by comprising the following steps:
the method comprises the following steps: carrying out femur trunk point cloud set P on the collected femur model point cloud set P shaft Obtaining a femur shaft point cloud set P shaft Relabeled point cloud set P upper And a relabeled point cloud set P lower (ii) a Relabeled point cloud set P upper Comprises a femoral head, a femoral neck and a femoral trochanter; relabeled point cloud set P lower Comprises a medial femoral condyle and a lateral femoral condyle;
step two: for relabeled point cloud set P upper Extracting femoral head Point cloud set P head ；
Step three: extracting femoral trochanter point cloud set P rotor And femur neck point cloud set P neck ；
Step four: extracting point cloud set P of medial condyle of femur intercondylar And lateral condyle area point cloud set P lateralcondylar 。
2. The method for medical semantic segmentation of a femoral point cloud model according to claim 1, wherein the first step comprises:
step 1a: segmenting the femur model point cloud set P into an initial point cloud set P upper And P lower ；
Step 1b: extracting point cloud set P upper Point cloud collection P of upper segment of middle femoral shaft upper-shaft ；
Step 1c: extracting point cloud set P lower Lower point cloud set P of middle femoral shaft lower-shaft ；
Step 1d: merging point cloud set P upper-shaft And a cloud set P lower-shaft To complete the femur trunk point cloud set P shaft And (4) extracting.
3. The method for medical semantic segmentation of a femoral point cloud model according to claim 2, wherein in step 1a, an initial point cloud set P is obtained upper And P lower The segmentation method comprises the following steps: constructing a femur point cloud model axial alignment bounding box AABB, and setting the axial direction of a femur model point cloud set P as a Y coordinate axis; counting the maximum value Ymax and the minimum value Ymin of the femur model point cloud set P in the Y coordinate axis direction, and centering the point in the Y coordinate axis direction
Dividing the femur point cloud model into two point cloud sets P upper And P lower ；
Wherein an initial set of point clouds P upper Comprises a femoral shaft upper segment, a femoral head, a femoral neck and a femoral trochanter, and an initial point cloud set P lower Comprises a femoral shaft lower segment, a femoral medial condyle and a femoral lateral condyle.
4. The method of claim 3, wherein in step 1b, an initial set of point clouds P is extracted upper The method for collecting the femur trunk point cloud comprises the following steps: according to the stock shaft characterization index CIS, an initial point cloud set P is subjected to upper Performing binary iteration segmentation;
the method comprises the following steps of: firstly, a point cloud slice with the thickness of delta at the Y coordinate axis Y = Y is obtained, namely [ Y-delta/2, Y + delta/2]Interior point cloud slices, then projecting the point cloud slices to plane Y = Y; then, calculating the slice projection point cloud centroid
Wherein P is i ＝(x i ,y i ,z i )，i＝1,2,…,n，P i Point coordinates, x, for point cloud slices i ,y i ,z i Are respectively a point P i The coordinate value of XYZ axis of (i) is the number of points; finally, point P is calculated i From the point cloud centroid P c Maximum value of distance D max And a minimum value D min Then CIS = D max -D min ；
The binary iterative segmentation is based on the Y-axis direction of the femur point cloud model
Whether CIS at the position (i =1,2,3, \8230;, n) is larger than a specified threshold value or not is judged, and a femoral shaft point cloud subset, namely a femoral shaft upper segment point cloud set P is extracted upper-shaft I is the number of binary iterative segmentations, yi is the initial point cloud set P upper Halving and dividing the maximum value Ymax and the minimum value Ymin;
wherein, the threshold value range is as follows: 5-8mm;
repeatedly executing binary iteration segmentation until the binary iteration cutting point is smaller than the point cloud slice thickness delta and the point cloud set P on the upper segment of the femoral shaft upper-shaft Is extracted, at which point a set of relabeled points P upper Comprises femoral head, femoral neck and femoral trochanter;
in step 1c, an initial point cloud P is collected lower Rotating pi around the Z axis, executing the step 1b again, and extracting the lower point cloud set P of the femoral shaft lower-shaft (ii) a At this point, the relabeled point cloud P lower Including the medial and lateral femoral condyles.
5. The medical semantic segmentation method of the femoral point cloud model according to claim 4, wherein the second step comprises:
step 2a: relabeling-based point cloud P upper Constructing a point cloud fitting sphere, and extracting a femoral head initial point cloud set P in ；
And step 2b: extracting the initial point cloud set P of the femoral head in And judging the type of the point cloud boundary;
and step 2c: executing breadth-first k adjacent region growing algorithm to extract femoral head point cloud set P head 。
6. The method for medical semantic segmentation of a femoral point cloud model according to claim 5, wherein in step 2a, the relabeled point cloud set P upper In and out of selectionCoplanar point P 0 、P 1 、P 2 And P 3 Construction of the center of sphere as (x) 0 ,y 0 ,z 0 ) An initial fitting sphere S with radius R, whose equation is (x-x) 0 ) 2 +(y-y 0 ) 2 +(z-z 0 ) 2 ＝R 2 (ii) a Then, a re-labeled point cloud set P is calculated upper Point p (x) of i ,y i ,z i ) Distance to the sphere S
If d is<Dividing threshold value, the value range of the dividing threshold value is 1-2 mm, then marking the point as a spherical fitting point P a ；
Repeating the above steps n times, if n is not less than 1800, then the point cloud P is marked again upper Point cloud set P of upper femoral head point set which is spherically fitted and relabeled upper The fitting spherical point cloud set is marked as the femoral head initial point cloud set P in The rest point cloud sets are marked as P out ；
In step 2b, the point cloud boundary is divided into an adjacent boundary AD and an internal boundary ID;
the abutment boundary AD is the boundary between the femoral head and the femoral neck;
the internal boundary ID is the femoral head point cloud not extracted by step 2a, which is caused by the local concavity of the femoral head surface;
adjacent boundary AD belongs to point cloud set P in And with P out The extraction basis of the adjacent point sets is as follows: order to
N k (P a ) Is a group of general formula with P a K is 15 if N is the point set formed by k points nearest to each other k 
Then P is a As a set of point clouds P in The boundary point of (2); the boundary point set is an adjacent boundary AD;
the method for judging the internal boundary ID comprises the following steps: as can be seen from the femoral head convex-concave region, if the number of boundary points of the convex-concave region is less than 18, the boundary is the inner boundary ID.
In step 2c, the breadth-first k-nearest neighbor region growing algorithm is as follows:
from point P e Starting from the point of view, extracting k neighbor point sets N k (P e ) And adding it to the initial point cloud set P of femoral head in Then recursively extracts N k (P e ) K nearest neighbor set N of the most distant point P' from the center in the intersection of (A) k (P') up to
Taking the internal boundary ID in the step 2b as a clue, repeatedly executing breadth-first k neighbor region growing, extracting femoral head point cloud inside the internal boundary ID and not extracted in the step 2a, and supplementing the femoral head point cloud to the femoral head initial point cloud set P in And completing the femoral head point cloud set P until the femoral head point cloud which is not extracted does not exist any more head And (6) dividing.
7. The medical semantic segmentation method of the femoral point cloud model according to claim 6, wherein the third step comprises:
step 3a: extracting femoral trochanter region characteristic line cluster P clue ；
And step 3b: extracting a femoral trochanter point set P by taking the characteristic line as a clue rotor 。
8. The medical semantic segmentation method for the femoral point cloud model according to claim 7, wherein the feature line cluster P clue The device consists of a horizontal characteristic line and a vertical characteristic line;
the extraction process of the horizontal characteristic line and the vertical characteristic line is as follows: firstly, making a point cloud slice which is vertical to an X axis and has the thickness of delta; then, extracting a central axis of the point cloud slice; finally, according to the femoral trochanteric features, taking the maximum point in the Y direction of the central axis of the point cloud slice as a segmentation point, and obtaining a current position feature line;
taking the current position characteristic line as a clue, executing an breadth-first k neighbor region growing algorithm in the step 2c, and completing the extraction of the femoral trochanter point cloud set; obtaining femoral trochanter point cloud set P rotor And femur neck point cloud set P neck 。
9. The medical semantic segmentation method of the femoral point cloud model according to claim 8, wherein the step of obtaining the segmentation feature points of the medial femoral condyle and the lateral femoral condyle comprises the steps of: firstly, making a point cloud slice in the vertical direction; then, solving a midpoint value mid in the X-axis direction of a point cloud slice characteristic line, wherein the point cloud slice characteristic line is the central axis of the point cloud slice; finally, the minimum value of the inner Y axis of the interval mid-5 and x are less than or equal to mid +5 is obtained, and the minimum value is the inner and outer condyle segmentation characteristic point P b ；
According to the segmentation feature point P b1 (x 1 ,y 1 ,z 1 )、P b2 (x 2 ,y 2 ,z 2 )、P b3 (x 3 ,y 3 ,z 3 ) Determining a plane Π: ax + By + Cz + D =0, computing a set P of relabeled point clouds lower Point P of b (x i ,y i ,z i ) Pi distance to plane
Wherein A, B, C and D are plane equation coefficients; setting a threshold value to be approximately equal to 0, and extracting a point cloud set P marked again lower The inner and outer condyles are divided into cloud sets of point to form dividing characteristic lines;
using the segmented feature lines to re-label the point cloud set P lower Segmented into femoral medial condyle point cloud set P intercondylar And lateral condyle area point cloud set P lateralcondylar 。
10. Use of a method of medical semantic segmentation of a femoral point cloud model according to any one of claims 1 to 9 for parametric measurement of the femur, personalized implant design and surgical planning.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210953666.3A CN115272615B (en) | 2022-08-10 | 2022-08-10 | Femur point cloud model medical semantic segmentation method and application thereof |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210953666.3A CN115272615B (en) | 2022-08-10 | 2022-08-10 | Femur point cloud model medical semantic segmentation method and application thereof |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115272615A true CN115272615A (en) | 2022-11-01 |
| CN115272615B CN115272615B (en) | 2023-09-08 |
Family
ID=83751478
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210953666.3A Active CN115272615B (en) | 2022-08-10 | 2022-08-10 | Femur point cloud model medical semantic segmentation method and application thereof |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115272615B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117257453A (en) * | 2023-07-20 | 2023-12-22 | 中国科学院自动化研究所 | Preoperative planning system for elbow joint replacement operation |
Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101839701A (en) * | 2010-05-27 | 2010-09-22 | 中国科学院自动化研究所 | Method for automatically partitioning tree point cloud data |
| US20130211531A1 (en) * | 2001-05-25 | 2013-08-15 | Conformis, Inc. | Patient-adapted and improved articular implants, designs and related guide tools |
| CN103632371A (en) * | 2013-12-06 | 2014-03-12 | 河海大学常州校区 | Compatibility mesh segmentation based skeleton parameter computation method |
| CN105139442A (en) * | 2015-07-23 | 2015-12-09 | 昆明医科大学第一附属医院 | Method for establishing human knee joint three-dimensional simulation model in combination with CT (Computed Tomography) and MRI (Magnetic Resonance Imaging) |
| CN205831965U (en) * | 2016-04-28 | 2016-12-28 | 华南理工大学 | A kind of personalized femoral prosthesis |
| CN109636896A (en) * | 2018-11-21 | 2019-04-16 | 河海大学常州校区 | Skeleton and its bone plate three-dimension interaction methods of exhibiting based on WebGL |
| CN110811829A (en) * | 2019-11-06 | 2020-02-21 | 中国人民解放军总医院第四医学中心 | Construction method and system based on femoral rotation axis and varus analysis model |
| CN111161267A (en) * | 2019-12-09 | 2020-05-15 | 西安工程大学 | Segmentation method of three-dimensional point cloud model |
| CN113409301A (en) * | 2021-07-12 | 2021-09-17 | 上海精劢医疗科技有限公司 | Point cloud segmentation-based femoral neck registration method, system and medium |
| CN113902937A (en) * | 2021-09-23 | 2022-01-07 | 中国科学院国家空间科学中心 | Self-adaptive target detection method and system for TOF point cloud |
-
2022
- 2022-08-10 CN CN202210953666.3A patent/CN115272615B/en active Active
Patent Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20130211531A1 (en) * | 2001-05-25 | 2013-08-15 | Conformis, Inc. | Patient-adapted and improved articular implants, designs and related guide tools |
| CN101839701A (en) * | 2010-05-27 | 2010-09-22 | 中国科学院自动化研究所 | Method for automatically partitioning tree point cloud data |
| CN103632371A (en) * | 2013-12-06 | 2014-03-12 | 河海大学常州校区 | Compatibility mesh segmentation based skeleton parameter computation method |
| CN105139442A (en) * | 2015-07-23 | 2015-12-09 | 昆明医科大学第一附属医院 | Method for establishing human knee joint three-dimensional simulation model in combination with CT (Computed Tomography) and MRI (Magnetic Resonance Imaging) |
| CN205831965U (en) * | 2016-04-28 | 2016-12-28 | 华南理工大学 | A kind of personalized femoral prosthesis |
| CN109636896A (en) * | 2018-11-21 | 2019-04-16 | 河海大学常州校区 | Skeleton and its bone plate three-dimension interaction methods of exhibiting based on WebGL |
| CN110811829A (en) * | 2019-11-06 | 2020-02-21 | 中国人民解放军总医院第四医学中心 | Construction method and system based on femoral rotation axis and varus analysis model |
| CN111161267A (en) * | 2019-12-09 | 2020-05-15 | 西安工程大学 | Segmentation method of three-dimensional point cloud model |
| CN113409301A (en) * | 2021-07-12 | 2021-09-17 | 上海精劢医疗科技有限公司 | Point cloud segmentation-based femoral neck registration method, system and medium |
| CN113902937A (en) * | 2021-09-23 | 2022-01-07 | 中国科学院国家空间科学中心 | Self-adaptive target detection method and system for TOF point cloud |
Non-Patent Citations (2)
| Title |
|---|
| DAVID A. J. WILSON, ET AL: "Validation of three-dimensional models of the distal femur created from surgical navigation point cloud data for intraoperative and postoperative analysis of total knee arthroplasty", 《INT J CARS》, pages 1 - 9 * |
| 吴云燕等: "基于兼容性网格分割的骨骼参数计算", 《计算机辅助设计与图形学学报》, pages 1134 - 1143 * |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117257453A (en) * | 2023-07-20 | 2023-12-22 | 中国科学院自动化研究所 | Preoperative planning system for elbow joint replacement operation |
Also Published As
| Publication number | Publication date |
|---|---|
| CN115272615B (en) | 2023-09-08 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| JP7384866B2 (en) | Bone reconstruction and orthopedic implants | |
| Lamecker et al. | A 3D statistical shape model of the pelvic bone for segmentation | |
| CN104091365B (en) | Towards the acetabular bone tissue model reconstruction method of serializing hip joint CT images | |
| US8200355B2 (en) | Method for manufacturing artificial implants | |
| CN107274389B (en) | Femur and acetabulum anatomical parameter obtaining method based on CT three-dimensional sequence image | |
| US20060094951A1 (en) | Computer-aided-design of skeletal implants | |
| CN108597017A (en) | A kind of textured bone template construction method based on measurement parameter | |
| US20220039959A1 (en) | Custom hip design and insertability analysis | |
| Jimenez-Carretero et al. | Optimal multiresolution 3D level-set method for liver segmentation incorporating local curvature constraints | |
| CN115272615A (en) | Medical semantic segmentation method for femoral point cloud model and application of medical semantic segmentation method | |
| CN111292306A (en) | Knee joint CT and MR image fusion method and device | |
| CN111724389A (en) | Hip joint CT image segmentation method, device, storage medium and computer equipment | |
| CN117530772B (en) | Method, device, medium and equipment for processing image before shoulder joint replacement operation | |
| CN111402230B (en) | Femoral medullary cavity morphological feature analysis method based on Monte Carlo method | |
| CN110811829B (en) | Construction method and system based on femoral rotation axis and varus analysis model | |
| CN115252233B (en) | Automatic planning method for acetabular cup in total hip arthroplasty based on deep learning | |
| CN112164075B (en) | Segmentation method for maxillary sinus membrane morphology change | |
| CN110766666B (en) | Skeleton fault outline large sample library generation method based on region segmentation and GAN model | |
| CN116934885A (en) | Lung segmentation method, device, electronic equipment and storage medium | |
| CN109145977B (en) | Bone damage type discrimination method based on naive Bayes | |
| CN109903390B (en) | Anatomical feature-based proximal femur averaged surface model construction method | |
| Song et al. | Computer-aided modeling and morphological analysis of hip joint | |
| Chen | Feature point extraction of femur surface based on reference object | |
| Sepasian et al. | Bone reposition planning for corrective surgery using statistical shape model: assessment of differential geometrical features | |
| Luan et al. | Point extraction from cross sections of fractured long bones for registration |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |