CN114565652A - Point cloud registration algorithm based on head features - Google Patents
Point cloud registration algorithm based on head features Download PDFInfo
- Publication number
- CN114565652A CN114565652A CN202210202661.7A CN202210202661A CN114565652A CN 114565652 A CN114565652 A CN 114565652A CN 202210202661 A CN202210202661 A CN 202210202661A CN 114565652 A CN114565652 A CN 114565652A
- Authority
- CN
- China
- Prior art keywords
- point cloud
- registration
- point
- head
- candidate
- 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.)
- Pending
Links
Images
Classifications
-
- 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
- G06T7/33—Determination of transform parameters for the alignment of images, i.e. image registration using feature-based methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T19/00—Manipulating 3D models or images for computer graphics
- G06T19/20—Editing of 3D images, e.g. changing shapes or colours, aligning objects or positioning parts
-
- 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
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30196—Human being; Person
- G06T2207/30201—Face
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2219/00—Indexing scheme for manipulating 3D models or images for computer graphics
- G06T2219/20—Indexing scheme for editing of 3D models
- G06T2219/2004—Aligning objects, relative positioning of parts
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2219/00—Indexing scheme for manipulating 3D models or images for computer graphics
- G06T2219/20—Indexing scheme for editing of 3D models
- G06T2219/2016—Rotation, translation, scaling
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Architecture (AREA)
- Computer Graphics (AREA)
- Computer Hardware Design (AREA)
- General Engineering & Computer Science (AREA)
- Software Systems (AREA)
- Magnetic Resonance Imaging Apparatus (AREA)
Abstract
With the development and application of computer imaging technology in medicine, medical imaging technology has become one of the most focused and rapidly developing fields in medical technology. The Magnetic Resonance Imaging (MRI) technology is a novel medical image examination technology developed by using a computer technology and an image reconstruction technology on the basis of finding a Magnetic resonance phenomenon in the field of physics. The head MRI images may be processed using brain storm software (braistorm) and a point cloud of the head surface generated. The OptiTrack is an optical system for realizing high-precision three-dimensional reconstruction of a mark point, the main hardware is an intelligent camera provided with a processor, a high-precision mark point processing algorithm is arranged in the intelligent camera, and three-dimensional information of a single mark point can be used for creating a rigid body and obtaining six-degree-of-freedom data. The invention provides a point cloud registration algorithm based on head features, which can complete high-precision head point cloud registration under the condition of only acquiring a small number of points of the head by using the point cloud features of the head, thereby reducing the waiting time of a patient and improving the treatment precision of an instrument.
Description
Technical Field
The invention relates to the field of computer vision, in particular to a point cloud registration algorithm based on head features.
Background
With the development and application of computer imaging technology in medicine, medical imaging technology has become one of the most focused and rapidly developing fields in medical technology. The Magnetic Resonance Imaging (MRI) technology is a novel medical image examination technology developed by using a computer technology and an image reconstruction technology on the basis of finding a Magnetic resonance phenomenon in the field of physics. The method has the advantages of safety, no wound, high resolution ratio of brain and other soft tissues, and multi-aspect multi-parameter imaging, and is widely applied to the field of brain science. The head MRI images may be processed using brain storm software (braistorm) and a point cloud of the head surface generated. The OptiTrack is an optical system for realizing high-precision three-dimensional reconstruction of a mark point, the main hardware is an intelligent camera provided with a processor, a high-precision mark point processing algorithm is arranged in the intelligent camera, and three-dimensional information of a single mark point can be used for creating a rigid body and obtaining six-degree-of-freedom data.
The iterative closest point algorithm (ICP) proposed by Besl and McKay in 1992 is currently the most widely used one of the point cloud registration algorithms. The disadvantages of this algorithm are, however: (1) the requirement on the initial position is high, a good initial posture needs to exist between two point clouds, otherwise iteration is not converged or a local optimal solution is trapped, and finally mismatching or non-convergence is caused. (2) The traditional ICP algorithm does not eliminate data with large noise in an iteration process, so that some points with large noise can have large influence on a final registration result. (3) When a subset relation exists between two groups of point clouds, the traditional ICP algorithm may cause the situation that iteration results are not converged. Therefore, the conventional ICP algorithm cannot meet the requirements of head point cloud registration.
Disclosure of Invention
In order to overcome the defects in the prior art, the invention provides a point cloud registration algorithm based on head features, and the high-precision head point cloud registration can be completed under the condition of only acquiring a small number of points of the head by utilizing the point cloud features of the head, so that the waiting time of a patient is reduced, and the instrument treatment precision is improved. The method specifically comprises the following steps:
step 1, acquiring source point cloud and target point cloud
Processing the head MRI image using a braistorm software and generating a source point cloud; acquiring three-dimensional point cloud data under a world coordinate system by using an OptiTrack camera and a pen with a marker ball to generate target point cloud;
step 2, carrying out coarse registration on the source point cloud and the target point cloud
Respectively taking a nasion point NAS, a left ear anterior point LPA and a right ear anterior point RPA in the source point cloud and the target point cloud, and performing coarse registration to obtain coarse registration results of the source point cloud and the target point cloud: rotation matrix RcAnd a translation vector tc;
Step 3, self-adaptive threshold screening;
step 4, performing fine registration by using the coarse registration result and the screened threshold value to obtain a fine registration result;
step 5, optimizing a fine registration result to minimize the cost;
and 6, calculating a final registration result.
2. A head feature based point cloud registration algorithm as claimed in claim 1 wherein: the step 3 specifically includes:
3.1) projecting the target point cloud by using the coarse registration result to obtain coarse registration point cloud Xc;
3.2) calculating XcThe mutual distance of NAS, LPA and RPA in the point cloud is marked as d1、d2And d3;
3.3) traversing the source point cloud, finding the sum X in the source point cloudcA set X consisting of a plurality of points with the shortest distances among NAS, LPA and RPA in the point cloudcNASs、XcLPAsAnd XcRPAs;
3.4) setting the screening range and the screening interval of the threshold value, and starting the screening threshold value: traverse XcNASs、XcLPAsAnd XcRPAsIs provided with XcNASsAny point in is xcNAS,XcLPAsAny point in is xcLPA,XcRPAsAny point in is xcRPASeparately calculate xcNASAnd xcLPAIs a distance dc1Calculating xcNASAnd xcRPAA distance of dc2Calculating xcLPAAnd xcRPAA distance of dc3(ii) a If | d1-dc1|、|d2-dc2I and | d3-dc3All values of | are less than the threshold, then x is considered to becNAS、xcLPAAnd xcRPAA set of candidate point pairs may be formed, with all candidate point pairs below the threshold being grouped into sets:
if set XcsetIf the number of the included elements is larger than a certain number, recording the threshold value as an alternative threshold value, after recording all alternative threshold values, judging whether the number of the alternative threshold values is larger than 0, if so, selecting the minimum alternative threshold value as a screening result, marking the threshold value as thrd, and executing a step 4; if not, the program is ended and the registration fails.
3. A head feature based point cloud registration algorithm as claimed in claim 2, wherein: the step 4 specifically includes:
4.1) selecting candidate point pairs under threshold thrd, and respectively calculating XcsetCandidate point pairs in the collection and coarse registration point cloud XcRotation matrix R of NAS, LPA and RPAcfAnd a translation vector tcfTo form a set Rt:
4.2) taking an arbitrary rotation matrix within the set RtAnd translation vectorBy passingObtaining candidate fine registration point cloud
4.3) traversing candidate fine registration point cloudAll points in the interior, computing and target point cloud XtIf the shortest distance is less than 3mm, the point is considered as an inner point after registration;
4.4) counting all candidate fine registration point cloudsSelecting the candidate fine registration point cloud with the maximum number of interior pointsCorresponding rotation matrix RfAnd a translation vector tfAs a result of the fine registration.
4. A head feature based point cloud registration algorithm as claimed in claim 3 wherein: the step 5 specifically includes:
5.1) point cloud X obtained by coarse registration transformationcRotation matrix R obtained by fine registrationfAnd a translation vector tfCarry out Xf=RfXc+tfTransforming to obtain fine registration point cloud Xf;
5.2) traversing the fine registration point cloud XfAnd a target point cloud XtIf the distance between the point in the precision registration point cloud and the point in the target point cloud is less than 3mm, the two points are considered as matching point pairs, and a matching point pair set between the two groups of point clouds is obtained through calculation;
5.3) calculating a fine registration point cloud X by using an ICP point cloud registration algorithmfAnd an objectPoint cloud XtR of the rotation matrix RsAnd a translation vector tsThe cost is minimized.
5. A head feature based point cloud registration algorithm as claimed in claim 1 wherein: the step 6 specifically includes: using a rotation matrix Rc、RfAnd RsAnd a translation vector tc、tfAnd tsComputing a rotation matrix R of the final registration resultnAnd a translation vector tn:
Rn=(RsRfRc)T
Make the source point cloud XsAnd a target point cloud XtSatisfy Xt=RnXs+tnAnd obtaining a final registration result.
The invention has the beneficial effects that:
(1) aiming at the problem that the traditional point cloud registration algorithm is invalid when the two groups of point clouds are low in overlapping degree and the set characteristics of the overlapping parts are not obvious, the invention can effectively realize the registration of the two groups of head point clouds with low overlapping degree and the set characteristics of the overlapping parts are not obvious by combining the head characteristics.
(2) By using three points of NAS, LPA and RPA of the human head as constraints, when the difference between the number of points of the source point cloud and the number of points of the target point cloud is large, and the target point cloud is a subset of the source point cloud, an accurate registration result can still be obtained.
(3) By using the NAS of the human head, the LPA and the RPA as constraints, error points of a registration point set can be effectively screened and eliminated, and the registration precision and robustness are improved.
Drawings
Fig. 1 is a flow chart of a point cloud registration algorithm based on head features according to an embodiment of the present invention.
Fig. 2 is a source cloud obtained by processing head MRI images using the braistorm software.
Fig. 3 is a cloud of target points obtained by capturing three-dimensional points in a world coordinate system using an OptiTrack camera and a pen with a marker ball.
Fig. 4 is a schematic diagram of a point cloud registration result of a conventional ICP algorithm.
FIG. 5 is a schematic diagram illustrating the result of point cloud registration according to the present invention.
Detailed Description
The following description of the embodiments of the present invention is provided in order to better understand the present invention for those skilled in the art with reference to the accompanying drawings. It is to be noted that the following description is only for explaining the present invention and is not intended to limit the present invention.
A point cloud registration algorithm based on head features, comprising the steps of:
step 1, acquiring a source point cloud and a target point cloud.
1.1) processing the head MRI image using the Brainstarm software and generating a point cloud of the head surface, called source point cloud, using XsTo indicate.
Wherein the content of the first and second substances,being any point of the source point cloud, the braistorm software can compute the coarse locations of the NAS, LPA, and RPA simultaneously.
1.2) using an OptiTrack camera and a positioning pen with a marker ball to collect head point cloud data under a world coordinate system, wherein the positions of the front three points are a Nasion point (NAS), a Left anterior auricular point (LPA) and a Right anterior auricular point (RPA), and the positions of the subsequent points are not required. This portion of the point cloud is called the target point cloud, using XtTo indicate.
And 2, performing coarse registration on the source point cloud and the target point cloud. Calculating NAS, LPA and RPA by using Brainsterm software as source point clouds, taking three points of the source point clouds, namely NAS, LPA and RPA, and setting coordinates of the three pointsUsing x separatelysNAS、xsLPAAnd xsRPATo indicate. Taking the first three points NAS, LPA and RPA of the target point cloud, setting the coordinates of the three points to respectively use xtNAS、xtLPAAnd xtRPATo indicate. Carrying out coarse registration by using three points of the source point cloud and the target point cloud to obtain a rotation matrix R between the source point cloud and the target point cloudcAnd a translation vector tc。
The steps of three-point registration are as follows: three points x using a source point cloudsNAS、xsLPAAnd xsRPACalculating the direction vectors of the x axis, the y axis and the z axis of the source point cloud, wherein the expression is as follows:
in the above formula, the expression represents a vector dot product, and the expression represents a vector cross product, obtained as aboveAndare all unit column vectors of 3 x 1, and thus the rotation matrix R of the source cloud with respect to the world coordinate systemsCan be expressed as:
front three points x using a target point cloudtNAS、xtLPAAnd xtRPACalculating the direction vector of the x-axis, y-axis and z-axis of the target point cloudAndthe rotation matrix R of the target point cloud with respect to the world coordinate systemtCan be expressed as:
the final three-point registration result is calculated using the following expression:
tc=xsNAS-RcxtNAS。
and 3, self-adapting to screen threshold values.
3.1) Point cloud X of the target PointtRotation matrix R obtained by coarse registrationcAnd a translation vector tcAnd (3) carrying out transformation:
Xc=RcXt+tc,
wherein, XcRepresenting a target point cloud XtAnd point cloud obtained by coarse registration transformation of the points.
3.2) calculating XcThe NAS and LPA distances in the point cloud are recorded as d1(ii) a Calculating XcThe distance between NAS and RPA in the point cloud is recorded as d2(ii) a Calculating XcThe distance between LPA and RPA in the point cloud is recorded as d3。
3.3) traversal of the Source Point cloud XsCalculating a source point cloud XsAnd coarse registration transformation point cloud XcUsing X as a candidate point set of the actual NAS, the 30 points with the nearest middle NAS distance are usedcNASsTo represent; computing a source point cloud XsAnd coarse registration transformation point cloud XcUsing X as the candidate point set of the actual LPA, the 30 points with the closest middle LPA distancecLPAsTo represent; computing a source point cloud XsAnd coarse registration transformation point cloud XcUsing 30 points with the nearest RPA as a candidate point set of the actual RPA and using XcRPAsTo indicate.
And 3.4) setting a screening range and a screening interval of the threshold value, and screening the minimum threshold value meeting the conditions. Traverse XcNASs、XcLPAsAnd XcRPAsIs provided with XcNASsAny point in is xcNAS,XcLPAsAt any point in itIs xcLPA,XcRPAsAny point in is xcRPASeparately calculate xcNASAnd xcLPAA distance of dc1Calculating xcNASAnd xcRPAA distance of dc2Calculating xcLPAAnd xcRPAA distance of dc3. If | d1-dc1|、|d2-dc2I and | d3-dc3All values of | are less than the threshold, then consider xcNAS, xcLPA and xcRPAA set of candidate point pairs may be formed. All candidate point pairs below the threshold may form a set:
if set XcsetIs greater than 10, the threshold is recorded as an alternative threshold. After recording all the alternative threshold values, judging whether the number of the alternative threshold values is larger than 0, if so, selecting the minimum alternative threshold value as a screening result, marking the threshold value as thrd, and executing a step 4; if not, the program is ended and the registration fails.
And 4, performing fine registration by using the coarse registration result and the screened threshold value.
4.1) selecting candidate point pairs under threshold thrd, and respectively calculating X by using a three-point registration algorithmcsetCandidate point pairs in the set and a coarse registration transformation point cloud XcOf NAS, LPA and RPAcfAnd a translation vector tcfThese rotation matrices RcfAnd a translation vector tcfA set Rt can be formed:
4.2) taking any set of rotation matrices in the set RtAnd translation vectorUsing a rotation matrixAnd translation vectorCalculating a candidate fine registration result:
wherein the content of the first and second substances,is the coarse registration point cloud Xtc passing through the candidate rotation matrixAnd translation vectorAnd transforming the obtained candidate fine registration point cloud.
4.3) traversing candidate fine registration point cloudAll points in the interior, calculating a target point cloud XtAnd the closest distance to the point, and if the closest distance is less than 3mm, the point is considered as the registered inner point.
N={n1,n2,…,ni}
set the maximum value Max (N) N in the set NjThen the value corresponds to the rotation matrixAnd translation vectorIs the result of the fine registration.
Wherein R isfIs the rotation matrix obtained by fine registration, tfIs the translation vector obtained by the fine registration.
And 5, optimizing a fine registration result.
5.1) point cloud X obtained by coarse registration transformationcRotation matrix R obtained by fine registrationfAnd a translation vector tfAnd (3) carrying out transformation:
Xf=RfXc+tf
wherein, XfRepresenting a point cloud XcPoint cloud obtained by fine registration transformation of the points.
5.2) obtaining the precision registration point cloud XfInner random pointTraversing target point cloud XtCalculating the target point cloud XtAndthe closest point isIf fine registration is performed in the point cloudAnd in the target point cloudIf the distance is less than 3mm, the two points are considered as matching point pairs, and the point cloud X of the fine registration is traversedfA set of matching point pairs may be obtained.
The set of matching points can in turn be represented using P and P':
P={p1,p2,p3,…,pi},
P′={p′1,p′2,p′3,…,p′i},
wherein p isi∈Xf,p′i∈XtAnd p isiAnd p'iAre pairs of matching points.
And 5.3) for the pose estimation problem of the two groups of point clouds with known matching points, a good result can be obtained by using the traditional ICP algorithm. Solving for secondary X using SVD decomposition methodfConversion to XtOf (3) a rotation matrix RsAnd a translation vector tsThe cost J is minimized, and the expression of the cost function J is as follows:
and 6, calculating a final registration result. The rotation matrix R obtained by calculation using the previous stepc、RfAnd RsAnd a translation vector tc、tfAnd tsA rotation matrix R of the final registration result can be calculatednAnd a translation vector tn:
Rn=(RsRfRc)T,
Make the source point cloud XsAnd a target point cloud XtSatisfies the following relationship:
Xt=RnXs+tn。
the present invention has been described in detail above, but the description of the embodiments is only for the purpose of explaining the method of the present invention and the core idea thereof so as to facilitate those skilled in the art to understand the present invention, but it should be clear that the present invention is not limited to the scope of the embodiments, and it is obvious to those skilled in the art that various changes can be made therein without departing from the spirit and scope of the present invention defined and determined by the appended claims, and all the inventions utilizing the inventive concept are protected.
Claims (5)
1. A point cloud registration algorithm based on head features is characterized in that: the method comprises the following steps:
step 1, acquiring source point cloud and target point cloud
Processing the head MRI image using a braistorm software and generating a source point cloud; acquiring three-dimensional point cloud data under a world coordinate system by using an OptiTrack camera and a pen with a marker ball to generate target point cloud;
step 2, carrying out rough registration on the source point cloud and the target point cloud
Respectively taking a nasion point NAS, a left ear anterior point LPA and a right ear anterior point RPA in the source point cloud and the target point cloud, and performing coarse registration to obtain coarse registration results of the source point cloud and the target point cloud: rotation matrix RcAnd a translation vector tc;
Step 3, self-adaptive threshold screening;
step 4, performing fine registration by using the coarse registration result and the screened threshold value to obtain a fine registration result;
step 5, optimizing a fine registration result to minimize the cost;
and 6, calculating a final registration result.
2. A head feature based point cloud registration algorithm as claimed in claim 1 wherein: the step 3 specifically includes:
3.1) projecting the target point cloud by using the coarse registration result to obtain coarse registration point cloud Xc;
3.2) calculating XcThe mutual distance of NAS, LPA and RPA in the point cloud is marked as d1、d2And d3;
3.3) traversing the source point cloud, finding the sum X in the source point cloudcSet X formed by a plurality of points with the shortest distances among NAS, LPA and RPA in point cloudcNASs、XcLPAsAnd XcRPAs;
3.4) setting the screening range and the screening interval of the threshold value, and starting the screening threshold value: traverse XcNASs、XcLPAsAnd XcRPAsIs provided with XcNASsAny point in is xcNAS,XcLPAsAny point in is xcLPA,XcRPAsAny point in is xcRPASeparately calculate xcNASAnd xcLPAA distance of dc1Calculating xcNASAnd xcRPAA distance of dc2Calculating xcLPAAnd xcRPAA distance of dc3(ii) a If | d1-dc1|、|d2-dc2I and | d3-dc3All values of | are less than the threshold, then x is considered to becNAS、xcLPAAnd xcRPAA set of candidate point pairs may be formed, with all candidate point pairs below the threshold being grouped into sets:
if set XcsetIf the number of the included elements is larger than a certain number, recording the threshold value as an alternative threshold value, after recording all alternative threshold values, judging whether the number of the alternative threshold values is larger than 0, if so, selecting the minimum alternative threshold value as a screening result, marking the threshold value as thrd, and executing a step 4; if not, the program is ended and the registration fails.
3. A head feature based point cloud registration algorithm as claimed in claim 2, wherein: the step 4 specifically includes:
4.1) selecting candidate point pairs under threshold thrd, and respectively calculating XcsetCandidate point pairs in the collection and coarse registration point cloud XcOf NAS, LPA and RPAcfAnd a translation vector tcfTo form a set Rt:
4.2) taking an arbitrary rotation matrix within the set RtAnd translation vectorBy passingObtaining candidate fine registration point cloud
4.3) traversing candidate fine registration point cloudAll points in the interior, computing and target point cloud XtIf the shortest distance is less than 3mm, the point is considered as an inner point after registration;
4. A head feature based point cloud registration algorithm as claimed in claim 3 wherein: the step 5 specifically includes:
5.1) point cloud X obtained by coarse registration transformationcRotation matrix R obtained by fine registrationfAnd a translation vector tfCarry out Xf=RfXc+tfTransforming to obtain fine registration point cloud Xf;
5.2) traversing the fine registration point cloud XfAnd a target point cloud XtIf the distance between the point in the precision registration point cloud and the point in the target point cloud is less than 3mm, the two points are considered as matching point pairs, and a matching point pair set between the two groups of point clouds is obtained through calculation;
5.3) calculating a fine registration point cloud X by using an ICP point cloud registration algorithmfAnd a target point cloud XtR of the rotation matrix RsAnd a translation vector tsThe cost is minimized.
5. A head feature based point cloud registration algorithm as claimed in claim 1 wherein: the step 6 specifically comprises: using a rotation matrix Rc、RfAnd RsAnd a translation vector tc、tfAnd tsComputing a rotation matrix R of the final registration resultnAnd a translation vector tn:
Rn=(RsRfRc)T
Make the source point cloud XsAnd a target point cloud XtSatisfy Xt=RnXs+tnAnd obtaining a final registration result.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202210202661.7A CN114565652A (en) | 2022-03-02 | 2022-03-02 | Point cloud registration algorithm based on head features |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202210202661.7A CN114565652A (en) | 2022-03-02 | 2022-03-02 | Point cloud registration algorithm based on head features |
Publications (1)
Publication Number | Publication Date |
---|---|
CN114565652A true CN114565652A (en) | 2022-05-31 |
Family
ID=81718341
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202210202661.7A Pending CN114565652A (en) | 2022-03-02 | 2022-03-02 | Point cloud registration algorithm based on head features |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN114565652A (en) |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN115844408A (en) * | 2022-11-16 | 2023-03-28 | 北京昌平实验室 | Magnetometer positioning device, brain magnetic positioning system and positioning method |
CN116563561A (en) * | 2023-07-06 | 2023-08-08 | 北京优脑银河科技有限公司 | Point cloud feature extraction method, point cloud registration method and readable storage medium |
-
2022
- 2022-03-02 CN CN202210202661.7A patent/CN114565652A/en active Pending
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN115844408A (en) * | 2022-11-16 | 2023-03-28 | 北京昌平实验室 | Magnetometer positioning device, brain magnetic positioning system and positioning method |
CN116563561A (en) * | 2023-07-06 | 2023-08-08 | 北京优脑银河科技有限公司 | Point cloud feature extraction method, point cloud registration method and readable storage medium |
CN116563561B (en) * | 2023-07-06 | 2023-11-14 | 北京优脑银河科技有限公司 | Point cloud feature extraction method, point cloud registration method and readable storage medium |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN114565652A (en) | Point cloud registration algorithm based on head features | |
US10535136B2 (en) | Registration of a magnetic tracking system with an imaging device | |
JP5464799B2 (en) | Image processing apparatus, image processing method, and program | |
US8861891B2 (en) | Hierarchical atlas-based segmentation | |
CN109887071A (en) | A kind of 3D video image dendoscope system and three-dimensional rebuilding method | |
CN112614169B (en) | 2D/3D spine CT (computed tomography) level registration method based on deep learning network | |
JP2020067748A (en) | Image processing apparatus, image processing method, and program | |
CN115578320A (en) | Full-automatic space registration method and system for orthopedic surgery robot | |
CN112258494B (en) | Focal position determination method and device and electronic equipment | |
CN113143459B (en) | Navigation method and device for laparoscopic augmented reality operation and electronic equipment | |
JP4887491B2 (en) | MEDICAL IMAGE PROCESSING METHOD, DEVICE THEREOF, AND PROGRAM | |
CN116883471A (en) | Line structured light contact-point-free cloud registration method for chest and abdomen percutaneous puncture | |
JP7312026B2 (en) | Image processing device, image processing method and program | |
CN111161143A (en) | Optical positioning technology-assisted operation visual field panoramic stitching method | |
CN116612166A (en) | Registration fusion algorithm for multi-mode images | |
CN112472293B (en) | Registration method of preoperative three-dimensional image and intraoperative perspective image | |
JP2023027751A (en) | Medical image processing device and medical image processing method | |
JP2000353244A (en) | Method for obtaining basic matrix, method for restoring euclidean three-dimensional information and device therefor | |
CN115018890A (en) | Three-dimensional model registration method and system | |
Wen et al. | Non-local attention learning for medical image classification | |
CN115116113A (en) | Optical navigation method | |
JPH0515525A (en) | Image position correction method | |
Taibi et al. | A Toolbox For Calibrating Fundus Camera | |
CN114418920B (en) | Endoscope multi-focus image fusion method | |
CN117036758B (en) | Two-dimensional image target matching method, electronic device and storage medium |
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 |