CN107967678B - Bone destruction degree feature extraction system - Google Patents
Bone destruction degree feature extraction system Download PDFInfo
- Publication number
- CN107967678B CN107967678B CN201710846927.0A CN201710846927A CN107967678B CN 107967678 B CN107967678 B CN 107967678B CN 201710846927 A CN201710846927 A CN 201710846927A CN 107967678 B CN107967678 B CN 107967678B
- Authority
- CN
- China
- Prior art keywords
- model
- bone
- sample
- pathological
- module
- 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.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0012—Biomedical image inspection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/213—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods
- G06F18/2135—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods based on approximation criteria, e.g. principal component analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/30—Determination of transform parameters for the alignment of images, i.e. image registration
-
- 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
- G06T2207/10072—Tomographic images
- G06T2207/10081—Computed x-ray tomography [CT]
-
- 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)
- Computer Vision & Pattern Recognition (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Evolutionary Computation (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Evolutionary Biology (AREA)
- General Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
- Radiology & Medical Imaging (AREA)
- Quality & Reliability (AREA)
- Measuring And Recording Apparatus For Diagnosis (AREA)
- Apparatus For Radiation Diagnosis (AREA)
- Ultra Sonic Daignosis Equipment (AREA)
Abstract
The invention discloses a bone destruction degree characteristic extraction system, which comprises a joint statistical form model construction module, a pathological form model inosculation module and an abnormity detection and analysis module, wherein the joint statistical form model construction module is used for constructing a joint statistical form model, and specific form statistical characteristics are extracted for each piece of bone; the pathological form model inosculating module is used for finding a model example which is closest to a pathological sample in the joint statistical form model, limiting the model example within three times of standard deviation of an average form model, eliminating the influence of normal anatomical variation and highlighting pathological change; the abnormal detection and analysis module is used for obtaining target information, namely the degree of each bone osteogenesis and osteoclast, and quantitative osteoclast and osteogenesis degree evaluation is realized by calculating positive and negative distances of corresponding vertexes. Therefore, the bone fracture evaluation index is calculated quantitatively, and reference information is provided for the evaluation of the bone fracture position and degree of the patient.
Description
Technical Field
The invention relates to the technical field of medical treatment, in particular to a bone destruction degree characteristic extraction system.
Background
The rheumatic arthritis belongs to autoimmune diseases, is highly developed in middle-aged and elderly people, but can also be seen in other age groups. Rheumatic arthritis is one of chronic diseases, and long-term rheumatic arthritis can cause limb disability. Some rheumatoid arthritis patients develop chronic inflammation of the synovial capsule, resulting in destruction of articular cartilage and bone mass. The bone destruction can be regarded as imbalance of two processes of bone formation and bone fracture, and the bone formation and bone fracture processes are finely regulated and controlled in a normal human body, so that the bone stability is ensured. Osteoclastic strength is greater than that of osteogenesis in patients with rheumatoid arthritis, leading to changes in joint structure and ultimately to osteoarticular dysfunction.
Evaluation of the degree of bone destruction of a patient often relies on medical personnel observing X-ray-based images to determine the degree of bone destruction of the patient, however, manual evaluation requires an evaluator with high professional expertise and is susceptible to subjective impressions.
Therefore, it is necessary to design a bone destruction level feature extraction system to solve the above technical problems.
Disclosure of Invention
Aiming at the problems in the background art, the invention aims to provide a bone destruction degree characteristic extraction system, which is used for acquiring the bone structure characteristics of normal people and patients from X-ray tomography so as to construct a grid model, extracting the bone destruction degree characteristics of the patients according to the difference between the two models, quantitatively calculating a bone destruction evaluation index and providing reference information for the evaluation of the bone destruction position and degree of the patients.
The technical scheme of the invention is realized as follows: the bone destruction degree feature extraction system comprises a joint statistical morphology model construction module, a pathological morphology model inosculation module and an anomaly detection and analysis module, wherein the joint statistical morphology model construction module: the method is used for constructing a joint statistical morphology model, and comprises the steps of firstly drawing a hierarchical system diagram of the skeleton structure of the whole body and depicting the connection mode among different skeletons; then acquiring a skeleton X-ray tomography image of a normal person to perform global rigid registration, and registering the whole reference grid model to an unmarked sample grid model; then, carrying out iterative registration on a single bone and the bones under the level system diagram by using an ICP algorithm; limiting the transformation parameters of the lower bone in a vertex subset in a truncated cone projected from the upper end; then, inverse transformation is used in the direction of an inverse level system diagram, so that the sample grid model is registered in the original coordinate system of the sample grid model; matching the reference grid model with the sample grid model, and projecting the label from the reference grid model to the sample grid model; for the description of the bone structure form variability, a point distribution model based on principal component analysis is used, after the point correspondence of each sample is established, the average form and the principal component of the bone structure are calculated, and specific form statistical characteristics are extracted for each piece of bone in a joint statistical form model; the pathological form model inosculating module is as follows: the method is used for finding a model example closest to a pathological sample in a joint statistical morphological model, limiting the model example to be within three times of standard deviation of an average morphological model, eliminating the influence of normal anatomical variation and highlighting pathological change; the pathological form model inosculating is divided into two steps, the first step is joint registration, the structure of the joint registration is different from the construction of a joint statistical form model in that the object registered by a marked reference model is a pathological sample, the pathological sample is marked and then converted back to the original coordinate system through inverse transformation, and simultaneously, the vertex one-to-one corresponding relation between the sample and the model is created; secondly, iterative operation is carried out by using a non-rigid registration method, so that the distance between the vertex of the mesh model and the vertex of the corresponding sample mesh model is minimized, and the iterative method is an active shape model fitting algorithm; the anomaly detection and analysis module: the target information for acquisition is the degree of bone formation and osteoclasts of each bone, and the abnormality detection realizes quantitative osteoclast and osteoblast degree evaluation by calculating the positive and negative distances of corresponding vertexes.
In the above technical solution, the connection manner between the bones includes connection between a trunk bone and a limb bone.
In the above technical solution, the non-rigid registration is obtained by performing linear combination definition according to principal component definition calculated by principal component analysis in a model building process, where x ≈ x + Pb, where a vector b parameterizes a deformation condition and is limited within plus or minus three times a standard deviation of an average value, P is a feature value matrix, and x are points of the average model and corresponding sample points.
In the above technical solution, in the iterative operation, every time an iteration is performed, the point correspondence between the two points is recalculated until the number of iterations reaches a set threshold n, and a model example is obtained through the iteration, so that the point correspondence is closest to the sample under the constraint condition imposed on the vector b.
In the above technical solution, the threshold n is set to 100.
The bone destruction degree feature extraction system comprises a joint statistical form model construction module, a pathological form model matching module and an anomaly detection and analysis module, wherein the joint statistical form model construction module is used for constructing a joint statistical form model and extracting specific form statistical features for each piece of bone; the pathological form model inosculating module is used for finding a model example which is closest to a pathological sample in the joint statistical form model, limiting the model example within three times of standard deviation of an average form model, eliminating the influence of normal anatomical variation and highlighting pathological change; the anomaly detection and analysis module is used for acquiring target information, namely the degree of each bone osteogenesis and osteoclast, and quantitative osteoclast and osteogenesis degree evaluation is realized by calculating positive and negative distances of corresponding vertexes. Therefore, the bone fracture evaluation index is calculated quantitatively, and reference information is provided for the evaluation of the bone fracture position and degree of the patient.
Drawings
FIG. 1 is a schematic diagram illustrating a hierarchy of the present invention;
fig. 2 is a combination diagram of several grids according to an embodiment of the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
The bone destruction degree feature extraction system comprises a joint statistical morphology model construction module, a pathological morphology model inosculation module and an abnormality detection and analysis module, and the following is a detailed description of the modules.
(1) A joint statistical morphology model construction module:
the joint statistical morphology model (ASSM) comprises morphological statistical characteristics of bone structures, and the first step of ASSM construction is to draw a hierarchical map of the bone structures of the whole body, which depicts the connection modes between different bones, such as the femur, the patella, the tibia, the fibula and the talus. And after obtaining a skeleton X-ray tomography image of a normal person, carrying out global rigid registration, and registering the whole reference grid model to the unmarked sample grid model. An ICP algorithm is then used to iteratively register a single bone and the bones that follow its hierarchical map. The transformation parameters for the underlying bone are limited to a subset of vertices within a frustum projected from the superior end. After completion, the inverse transform needs to be used in the inverse hierarchical diagram direction so that the sample grid model completes registration within its own original coordinate system. Through the bone joint registration, matching of the reference grid model and the sample grid model is achieved, and the labels are projected from the reference grid model to the sample grid model.
The description of the morphological variability of the bone structure uses a point distribution model based on principal component analysis. After the point correspondence relationship of each sample is established, the average morphology and principal component of the bone structure are calculated, and specific morphological statistical characteristics are extracted for each piece of bone in the ASSM. Wherein the diagram of the stage system is shown in FIG. 1.
(2) Pathological form model anastomosis module:
the goal of the pathological morphology model fitting is to find a model instance in the ASSM such that this model is closest to the pathological sample. By limiting the model instances to within three standard deviations of the mean morphological model, the effects of normal anatomical variation are eliminated and pathological changes are highlighted.
The pathological morphology model anastomosis is divided into two steps, the first step is joint registration, the structure of the joint registration is the same as the construction step of the ASSM, and the only difference is that the object registered by the marked reference model in the step is a pathological sample, and the pathological sample is marked and then converted back into the original coordinate system through inverse transformation. This process also creates a one-to-one correspondence of vertices between samples and models. In the second step, iterative operations are performed using a non-rigid registration method, such that the distances between the vertices of the mesh model and the vertices of the corresponding sample mesh model are minimized, the iterative method being an Active Shape Model (ASM) fitting algorithm. And the non-rigid registration is obtained by performing linear combination definition according to the principal component definition calculated by principal component analysis in the model construction process:
x≈x+Pb
the vector b parameterizes the deformation, constrained to within plus or minus three standard deviations of the mean, P is the eigenvalue matrix, and x are the points of the mean model and the corresponding sample points. Every time an iteration is performed, the point correspondence between the two points is recalculated until the number of iterations reaches a threshold n, which is generally set to 100. An example of the model is iteratively obtained so that it is closest to the sample under the constraints imposed on the vector b.
(3) An anomaly detection and analysis module:
the target information to be acquired by the anomaly detection and analysis module is the degree of each bone osteogenesis and osteoclastogenesis, and the anomaly detection is quantitative evaluation of the osteoclastogenesis and the osteoblastogenesis degree by calculating the positive and negative distances of corresponding vertexes.
Calculating Euclidean distance between corresponding points of the pathological form model and the reference model, and setting m to be (m)1，m2，m3) The point is a point on the reference model, and n is (n)1，n2，n3) The point is a corresponding point on the pathological form model, and the calculation formula of the Euclidean distance without symbols is as follows:
in order to determine whether the local area on the pathomorphological model is dominated by bone formation or bone fracture, it is necessary to calculate the signed euclidean distance. The reference model and the pathological form model are both grid structure models formed by a large number of vertexes, and the orthogonal vector of the surface of any surface can be obtained by the cross product of two edge vectors. For any vertex, its orthogonal vector is averaged with the orthogonal limit of the faces adjacent to it. The signed euclidean distance is:
SE(m，n)＝E(m，n)(sgn(N·(m-n)))
if the Euclidean distance with the symbol is close to zero, the two models at the position are well matched; positive values indicate active osteogenic activity and negative values indicate active osteoclastic activity.
After the calculation of the signed euclidean distance of each vertex is completed, the abnormal region needs to be depicted by using a heat map, so that a strong evaluation index is provided. First a tolerance threshold needs to be defined to reflect the deviation of the registration and model fit. The tolerance threshold is obtained by calculating the average Euclidean distance between each sample model and the reference model in the original training set, and for each bone, the numerical value is calculated independently, so that the registration precision variation corresponding to different bones is explained. If the deviation value falls within three standard deviations of the mean deviation value of each bone, the normal variation is considered, and the deviation above the tolerance threshold is considered abnormal, and an abnormal surface area calculation is performed.
The mesh surfaces adjacent to the vertices of the abnormal area are then clustered together to form labeled patches representing bone or bone fractures. Each cell surface in the plot has the same sign, with a negative sign indicating osteoclasts, a zero value indicating normal, and a positive sign indicating osteogenesis. The decision process is by making a decision on vertices one by one. In order to avoid the influence of local irregularity on the judgment of the grid surface sign, a judgment rule base is required to be defined, the judgment process takes a certain vertex as the center, the triangle plane signs around the vertex correspond to other vertexes, the peripheral vertexes are taken as input in a clockwise sequence and are adjusted to be not influenced by rotation, and then corresponding rule items are called from the judgment rule base, so that the grid surface signs around the vertex can be defined. The bone surface area of the corresponding marked small block is obtained by summing the areas of the small triangles of each mesh. For triangle v1v2v3The area is as follows:
S＝1/2|(v2-v1)×(v3-v1)|
the overall percentage coverage was used to measure the degree of bone destruction, including percent osteogenesis versus percent broken bone, for each piece of bone. Connected together tiles with the same sign using a queue-based connected region labeling algorithm is referred to as a maximized same-sign connected region, and the tile surface area is obtained by summing the triangular areas forming the tiles. Calculating the equivalent diameter, wherein the area of a circle corresponding to the diameter is equal to the total area of the small blocks, and the calculation method comprises the following steps:
the following describes, in conjunction with a specific example, the rules for determining the signs of mesh planes from vertices: fig. 2 is a schematic diagram of a combination of a plurality of grids, which is illustrated as follows:
the light colored vertices in the graph represent active osteogenic activity and the dark colored vertices represent active osteoclastic activity; the triangle marked negative sign indicates that the lattice triangle is active for osteoclastic activity and the positive sign indicates that the lattice triangle is active for osteoblastic activity.
And when the vertex a is judged, the vertex a is a five-level vertex, five triangles are arranged on the periphery, the five vertexes correspond to the five vertexes, the vertex a is in a clockwise sequence of red, blue and blue, entries corresponding to the five-level vertex are read from the rule base, according to the entries, two triangle definition symbols formed by the vertex and the other three blue vertexes are positive, and the other three triangles are defined as 0.
When the b vertex is reached, the b vertex is also a five-level vertex, the peripheral vertexes of the b vertex are adjusted to be red, blue and blue after clockwise ordering, and according to the entries in the rule base, a triangle formed by the two red vertexes and the vertex a is defined as negative. In this way, the stack determines one triangle to be 0 when determining the vertex a, and determines it to be negative again according to the rule when determining the vertex b.
Compared with the prior art, the bone destruction degree characteristic extraction system has the following beneficial effects:
1. the previous estimation of the bone destruction degree of a rheumatoid arthritis patient by a clinician only depends on the manual evaluation of a medical image based on X-ray, the evaluation process has high requirements on the professional knowledge level of an evaluator, and the deviation of judgment is easily caused by the image of subjective factors. The bone destruction degree characteristic extraction system provided by the invention constructs a joint morphology statistical model (ASSM) and a pathological morphology model in a modeling mode, and calculates the Euclidean distance with symbols between corresponding points after matching, so that the bone destruction degree is quantized by means of a plurality of parameters including bone destruction area and equivalent amount of tenderness, an image of subjective judgment is planed, and the bone destruction degree is more objective and more accurate.
2. If only quantitative indexes of bone fracture degree are provided, only numerical estimation of the bone fracture degree can be obtained by the indexes, the bone fracture is difficult to be accurately positioned, and the severity of the bone fracture in each bone is difficult to be judged. The invention judges the sign of each grid surface according to the defined rule according to the signed Euclidean distance of each vertex, connects the homogenized grid triangular surfaces into small blocks to form the small blocks respectively representing the dominance of osteogenic activity and the small blocks representing the dominance of osteoclastic activity, and finally presents the activity degree of osteoclastic bone of each bone by using a thermal diagram, so that a user can more intuitively sense the local bone destruction degree of each bone.
3. Because of the individual differences, the morphology of bone structures often varies among individuals, and the identification of anatomical variation from bone destruction is a challenge, requiring differentiation between anatomical variation and bone destruction. The ASSM constructed by the present invention is not simply constructed by directly averaging the skeleton structure of a normal person, but is constructed by a plurality of samples of normal persons, and it is required that the structure of any one example of the ASSM can be found among the normal persons. When the pathological morphology models are matched, the nearest examples are searched in the ASSM for registration, and the structure with the deviation value within the range of three times of the positive standard deviation and the negative standard deviation is regarded as normal, so that the error judgment of normal anatomical variation is avoided.
The above description is only for the purpose of illustrating the preferred embodiments of the present invention and is not to be construed as limiting the invention, and any modifications, equivalents, improvements and the like that fall within the spirit and principle of the present invention are intended to be included therein.
Claims (5)
1. A bone destruction degree feature extraction system is characterized in that: comprises a joint statistical morphology model construction module, a pathological morphology model inosculation module and an abnormality detection and analysis module, wherein,
the joint statistical morphology model construction module: the method is used for constructing a joint statistical morphology model, and comprises the steps of firstly drawing a hierarchical system diagram of a whole body skeleton structure and depicting a connection mode among different skeletons; then obtaining a skeleton X-ray tomography image of a normal person to carry out global rigid registration, and registering the whole reference grid model to an unmarked sample grid model; then, carrying out iterative registration on a single bone and the bones under the level system diagram by using an ICP algorithm; limiting the transformation parameters of the lower bone in a vertex subset in a truncated cone projected from the upper end; then, inverse transformation is used in the direction of an inverse level system diagram, so that the sample grid model is registered in the original coordinate system of the sample grid model; matching the reference grid model with the sample grid model, and projecting the label from the reference grid model to the sample grid model; for the description of the bone structure form variability, a point distribution model based on principal component analysis is used, after the point correspondence of each sample is established, the average form and the principal component of the bone structure are calculated, and specific form statistical characteristics are extracted for each piece of bone in a joint statistical form model;
the pathological form model inosculating module is as follows: the method is used for finding a model example closest to a pathological sample in a joint statistical morphological model, limiting the model example to be within three times of standard deviation of an average morphological model, eliminating the influence of normal anatomical variation and highlighting pathological change; the pathological form model inosculating is divided into two steps, the first step is joint registration, the structure of the joint registration is different from the construction of the joint statistical form model in that the object registered by the marked reference model is a pathological sample, the pathological sample is marked and then converted back to the original coordinate system through inverse transformation, and simultaneously, the vertex one-to-one corresponding relation between the sample and the model is created; secondly, iterative operation is carried out by using a non-rigid registration method, so that the distance between the vertex of the mesh model and the vertex of the corresponding sample mesh model is minimized, and the iterative method is an active shape model fitting algorithm;
the anomaly detection and analysis module: the target information for acquisition is the degree of each bone osteogenesis and osteoclastogenesis, and the abnormality detection realizes quantitative evaluation of the osteoclastogenesis and the osteogenesis degree by calculating the positive and negative distances of corresponding vertexes.
2. The bone destruction level feature extraction system according to claim 1, wherein: the connections between the bones include connections between the trunk bone and the limbs bone.
3. The bone destruction level feature extraction system according to claim 1, wherein: the non-rigid registration is obtained by performing linear combination definition according to principal component definition calculated by principal component analysis in the model construction process, wherein x 'is approximately equal to x + Pb, the vector b parameterizes the deformation condition and is limited within plus and minus three times of standard deviation of the average value, P is a characteristic value matrix, x' is a point of the average model, and x is a corresponding sample point.
4. The bone destruction level feature extraction system according to claim 3, wherein: in the iterative operation of the pathological form model matching module, every time iteration is carried out, the corresponding situation of the point between the two is recalculated until the iteration frequency reaches a set threshold value n, and a model example is obtained through iteration, so that the point is closest to a sample under the constraint condition exerted on a vector b.
5. The bone destruction level feature extraction system according to claim 4, wherein: the threshold n is set to 100.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710846927.0A CN107967678B (en) | 2017-09-18 | 2017-09-18 | Bone destruction degree feature extraction system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710846927.0A CN107967678B (en) | 2017-09-18 | 2017-09-18 | Bone destruction degree feature extraction system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN107967678A CN107967678A (en) | 2018-04-27 |
| CN107967678B true CN107967678B (en) | 2021-11-02 |
Family
ID=61996641
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710846927.0A Active CN107967678B (en) | 2017-09-18 | 2017-09-18 | Bone destruction degree feature extraction system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107967678B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113658170A (en) * | 2021-08-30 | 2021-11-16 | 杭州柳叶刀机器人有限公司 | Joint registration point generation method and device, electronic equipment and storage medium |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104614200A (en) * | 2015-02-10 | 2015-05-13 | 唐佩福 | Computer CT and 3D printing technology-based method for precisely sampling any cancellous bone area in human femoral head |
| CN105303604A (en) * | 2015-10-19 | 2016-02-03 | 中国科学院软件研究所 | Measuring method and system for single-side osteal damage of human body |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| GB0325523D0 (en) * | 2003-10-31 | 2003-12-03 | Univ Aberdeen | Apparatus for predicting bone fracture risk |
| GB0806509D0 (en) * | 2008-04-10 | 2008-05-14 | Nordic Bioscience As | Vertebral fracture prediction |
-
2017
- 2017-09-18 CN CN201710846927.0A patent/CN107967678B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104614200A (en) * | 2015-02-10 | 2015-05-13 | 唐佩福 | Computer CT and 3D printing technology-based method for precisely sampling any cancellous bone area in human femoral head |
| CN105303604A (en) * | 2015-10-19 | 2016-02-03 | 中国科学院软件研究所 | Measuring method and system for single-side osteal damage of human body |
Also Published As
| Publication number | Publication date |
|---|---|
| CN107967678A (en) | 2018-04-27 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Subburaj et al. | Automated identification of anatomical landmarks on 3D bone models reconstructed from CT scan images | |
| US10318839B2 (en) | Method for automatic detection of anatomical landmarks in volumetric data | |
| Anas et al. | Automatic segmentation of wrist bones in CT using a statistical wrist shape $+ $ pose model | |
| CN109692041B (en) | Bone surface registration method for covering cartilage | |
| Liu et al. | NHBS-Net: A feature fusion attention network for ultrasound neonatal hip bone segmentation | |
| AU2020101643A4 (en) | Method for automatically constructing patient-specific anatomic model based on statistical shape model (SSM) | |
| CN111429417B (en) | Method for detecting rotation center of femoral head of child | |
| CN116650115A (en) | Orthopedic surgery navigation registration method based on UWB mark points | |
| CN107967678B (en) | Bone destruction degree feature extraction system | |
| Medrano-Gracia et al. | Large scale left ventricular shape atlas using automated model fitting to contours | |
| Phan et al. | Predicting anatomical landmarks and bone morphology of the femur using local region matching | |
| Liu et al. | Sparse intervertebral fence composition for 3D cervical vertebra segmentation | |
| CN114938995B (en) | Pelvic registration system and medical device for use in hip replacement surgery | |
| Kang et al. | Simultaneous pose estimation and patient-specific model reconstruction from single image using maximum penalized likelihood estimation (MPLE) | |
| Pei et al. | Superimposition of cone-beam computed tomography images by joint embedding | |
| Chen et al. | Deep-learning based surface region selection for deep inspiration breath hold (DIBH) monitoring in left breast cancer radiotherapy | |
| CN101071455A (en) | Body image automatic standardizing method | |
| Ma et al. | Optimal mass transport based brain morphometry for patients with congenital hand deformities | |
| Gómez et al. | A real-coded evolutionary algorithm-based registration approach for forensic identification using the radiographic comparison of frontal sinuses | |
| Zhang et al. | Performance analysis of active shape reconstruction of fractured, incomplete skulls | |
| Huang et al. | Multi-modal medical image registration based on gradient of mutual information and morphological Haar wavelet | |
| Zhang et al. | Intelligent measurement of spinal curvature using cascade gentle AdaBoost classifier and region-based DRLSE | |
| Sanchez et al. | Segmentation-guided coordinate regression for robust landmark detection on X-rays: application to automated assessment of lower limb alignment | |
| Zhang et al. | Automatic liver segmentation from CT scans based on a statistical shape model | |
| Shi et al. | Human brain mapping with conformal geometry and multivariate tensor-based morphometry |
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 |