CN108764089B - Human back spine midline recognition method based on depth camera - Google Patents
Human back spine midline recognition method based on depth camera Download PDFInfo
- Publication number
- CN108764089B CN108764089B CN201810479290.0A CN201810479290A CN108764089B CN 108764089 B CN108764089 B CN 108764089B CN 201810479290 A CN201810479290 A CN 201810479290A CN 108764089 B CN108764089 B CN 108764089B
- Authority
- CN
- China
- Prior art keywords
- curvature
- matrix
- spine
- midline
- coordinates
- 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
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/103—Static body considered as a whole, e.g. static pedestrian or occupant recognition
Abstract
The invention discloses a human back spine midline recognition method based on a depth camera, which comprises the steps of obtaining a human back depth image, gridding, curvaturing, extracting a back midline region, obtaining a keel point coordinate and a crotch point coordinate, and carrying out cubic spline interpolation fitting to obtain a human back spine midline. According to the method, the curvature processing is carried out on the obtained depth image of the back of the human body, the information of the back of the human body is extracted, and the extracted part of the central line of the spine is processed by adopting a cubic line interpolation fitting method, so that the complete central line of the spine of the back of the human body can be obtained, and the accuracy of the central line extraction of the spine is improved.
Description
Technical Field
The invention belongs to the technical field of machine vision, and particularly relates to a human back spine midline identification method based on a depth camera.
Background
Scoliosis, which can also be called as scoliosis, has the characteristic of multiple occurrence, and in recent years, the occurrence of scoliosis in teenagers is increasing year by year, which has a certain influence on the life and the future work of the teenagers. Spine detection has therefore found considerable widespread use. The currently common method of scoliosis is the scoliosis ruler, Adams forward stooping test. However, due to the disadvantages of low efficiency, high labor intensity, and high error liability of manual detection, scoliosis detection is gradually progressing from manual processing to computer-automated processing today, which is a continuous development of machine vision.
Extraction and fitting of the spine curve is crucial for a detection system. The traditional scoliosis extraction has the following disadvantages. Firstly, methods such as measuring the rotation angle of the trunk by using a scoliosis ruler, Adams forward bending test and the like have large workload, and when a large number of people are generally checked, manual detection becomes quite cumbersome, the efficiency is very low, and a doctor may cause erroneous judgment and misjudgment due to fatigue. Second, the X-rays used in scoliosis detection have a certain amount of radioactivity, which is very detrimental to the growth and development of adolescents. The innocent detection newly developed abroad is very high in price, so that the detection is difficult to be used by common patients in ordinary hospitals, and therefore, the development of innocent, reliable and efficient detection equipment can bring great improvement to the spine detection problem.
Disclosure of Invention
The invention aims to: in order to solve the problems in the prior art, the invention provides a human back spine midline recognition method based on a depth camera.
The technical scheme of the invention is as follows: a human back spine midline recognition method based on a depth camera comprises the following steps:
A. acquiring a depth image of the back of a human body by using a depth camera;
B. b, gridding the depth image of the back of the human body obtained in the step A to obtain a grid matrix;
C. b, carrying out curvature processing on the grid matrix obtained in the step B to obtain a curvature matrix;
D. c, extracting a central line region of the back according to the curvature matrix obtained in the step C;
E. and D, obtaining the coordinates of the keel points and the coordinates of the crotch points, and performing cubic spline interpolation fitting processing by combining the back midline region obtained in the step D to obtain the spine midline of the back of the human body.
Further, the step C performs curvature processing on the grid matrix obtained in the step B to obtain a curvature matrix, which specifically includes: and C, calculating a second derivative and a first derivative of the grid matrix obtained in the step B along the horizontal direction, and combining the obtained second derivative and the obtained first derivative by adopting a curvature formula to obtain a curvature matrix.
Further, the curvature formula is expressed as:
where k represents the curvature and y is the depth image obtained by the depth camera.
Further, the step C further includes setting a curvature value with a negative curvature in the curvature matrix to zero, setting a curvature value with a curvature greater than 1 in the curvature matrix to zero, and setting an isolated curvature value in the curvature matrix to zero.
Further, the step D extracts a back centerline region according to the curvature matrix obtained in the step C, specifically: carrying out corrosion operation on the curvature matrix, and dividing the curvature matrix into a plurality of characteristic areas; then removing isolated points and carrying out expansion operation; and extracting the maximum connected domain of the curvature matrix to obtain a back midline region.
Further, the step E obtains the coordinates of the keel points and the coordinates of the crotch points, and performs cubic spline interpolation fitting processing in combination with the back centerline region obtained in the step D to obtain the spine centerline of the back of the human body, which specifically comprises: obtaining a keel point coordinate and a crotch point coordinate and converting the keel point coordinate and the crotch point coordinate into a depth image coordinate; d, taking a median boundary value from the left boundary and the right boundary of the dorsal midline region obtained in the step D to obtain a partial spinal midline; and selecting the coordinates of the discrete points on the straight line connecting the coordinates of the keel points and the upper vertexes of the central lines of the partial spines and the coordinates of the discrete points on the straight line connecting the coordinates of the middle points of the connecting lines of the two crotch bone points and the lower vertexes of the central lines of the partial spines, and performing cubic spline interpolation fitting processing by combining the central lines of the partial spines to obtain the central line of the spine of the human back.
Further, the method comprises the following steps of selecting discrete point coordinates on a straight line connecting the coordinates of the keel points and the upper vertexes of the central lines of the partial spines, selecting discrete point coordinates on a straight line connecting the coordinates of the middle points of the connecting lines of the two crotch bone points and the lower vertexes of the central lines of the partial spines, and performing cubic spline interpolation fitting processing by combining the central lines of the partial spines to obtain the central line of the spine of the back of the human body, wherein the method specifically comprises the following steps:
s1, constructing a cubic spline interpolation fitting model;
s2, determining the boundary coordinates of the equation according to the connecting line midpoint between the keel point and the two crotch points, and constructing a tri-diagonal matrix equation set;
s3, solving the tri-diagonal matrix equation set constructed in the step S2 by adopting a catch-up method to obtain diagonal parameters;
s4, reconstructing a cubic equation among the points according to the diagonal parameters obtained in the step S3 to obtain a cubic spline difference coefficient;
s5, in each interval xi < x < xi +1, drawing out each equation reconstructed in the step S4 point by point to obtain a fitted spinal curve.
Further, the cubic spline interpolation fitting model is specifically expressed as:
Pi(x)＝ai+bi(x-xi)+ci(x-xi)2+di(x-xi)3
wherein, ai,bi,ci,diIs a cubic spline coefficient of difference, xiIs the x coordinate of the i-th discrete point, and x is the x coordinate of the generated fitted curve.
Further, the tri-diagonal matrix equation set is specifically expressed as:
wherein m isiDerivative of the fitted curve for the i-th segment, hiIs the y coordinate of the i-th discrete point, mi＝Pi′(xi)，hi＝xi+1-xi。
Further, the cubic spline difference coefficient is specifically expressed as:
the invention has the beneficial effects that: according to the method, the curvature processing is carried out on the obtained depth image of the back of the human body, the information of the back of the human body is extracted, and the extracted part of the central line of the spine is processed by adopting a cubic line interpolation fitting method, so that the complete central line of the spine of the back of the human body can be obtained, and the accuracy of the central line extraction of the spine is improved.
Drawings
FIG. 1 is a flow chart of a human back spinal midline recognition method based on a depth camera according to the invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.
Fig. 1 is a schematic flow chart of a method for recognizing a spinal midline of a back of a human body based on a depth camera according to the present invention. A human back spine midline recognition method based on a depth camera comprises the following steps:
A. acquiring a depth image of the back of a human body by using a depth camera;
B. b, gridding the depth image of the back of the human body obtained in the step A to obtain a grid matrix;
C. b, carrying out curvature processing on the grid matrix obtained in the step B to obtain a curvature matrix;
D. c, extracting a central line region of the back according to the curvature matrix obtained in the step C;
E. and D, obtaining the coordinates of the keel points and the coordinates of the crotch points, and performing cubic spline interpolation fitting processing by combining the back midline region obtained in the step D to obtain the spine midline of the back of the human body.
In an alternative embodiment of the present invention, the step a uses a depth camera to obtain the depth image of the back of the human body, and the human body adopts an upright standing posture with the coat removed.
In an optional embodiment of the present invention, in the step B, the human back depth image obtained in the step a is subjected to gridding processing by using a quickhull algorithm, so as to obtain a grid matrix with a size of 200 × 200.
In an optional embodiment of the present invention, in the step C, the grid matrix obtained in the step B is subjected to curvature processing to obtain a curvature matrix, and specifically, the curvature matrix is: and C, calculating a second derivative and a first derivative of the grid matrix obtained in the step B along the horizontal direction, and combining the obtained second derivative and the obtained first derivative by adopting a curvature formula to obtain a curvature matrix. The curvature formula here is expressed as:
where k represents curvature, y is depth image data obtained by the depth camera, y' is a first derivative of the depth image data, and y "is a second derivative of the depth image data.
After the curvature matrix is obtained, the curvature value with negative curvature in the curvature matrix is set to zero, the curvature value with curvature larger than 1 in the curvature matrix is set to zero, and the isolated curvature value in the curvature matrix and the curvature value without left, right, upper and lower parts in the curvature matrix are set to zero.
In an optional embodiment of the present invention, the step D extracts a central line region on the back according to the curvature matrix obtained in the step C, specifically: firstly, carrying out corrosion operation on a curvature matrix, and dividing the curvature matrix into a plurality of characteristic regions including a scapular region, a back midline region and a waist region; removing isolated points, setting the isolated points to zero, performing expansion operation, and removing burrs on each area; and finally, extracting the maximum connected domain of the curvature matrix, wherein the maximum connected domain is the back midline region.
In an optional embodiment of the present invention, the foregoing step obtains the bone-raising point coordinates and the hip point coordinates, and performs cubic spline interpolation fitting processing in combination with the back midline region obtained in step D to obtain a spine midline of the back of the human body, specifically: obtaining a keel point coordinate and a crotch point coordinate according to the three-dimensional color image, and respectively converting the keel point coordinate and the crotch point coordinate into a depth image coordinate; taking a median boundary value from the left and right boundaries of the dorsal midline region obtained in the step D to obtain a partial spinal midline; selecting discrete point coordinates on a straight line connecting the coordinates of the keel points and the upper vertexes of the central lines of the partial spines and discrete point coordinates on a straight line connecting the coordinates of the middle points of the connecting lines of the two crotch bone points and the lower vertexes of the central lines of the partial spines, and performing cubic spline interpolation fitting processing by combining the central lines of the partial spines to obtain the central lines of the spines of the human backs, wherein the method specifically comprises the following steps:
s1, constructing a cubic spline interpolation fitting model, which is specifically expressed as:
Pi(x)＝ai+bi(x-xi)+ci(x-xi)2+di(x-xi)3
wherein, ai,bi,ci,diIs a cubic spline coefficient of difference, xiExtracting the x coordinate of the middle line area of the back for the x coordinate of the ith discrete point, wherein x is the x coordinate of the generated fitting curve;
s2, determining the boundary coordinates of the equation according to the connecting line midpoint between the keel point and the two crotch points, and constructing a tri-diagonal matrix equation set, wherein the specific expression is as follows:
wherein m isiDerivative of the fitted curve for the i-th segment, hiIs the y coordinate of the i-th discrete point, mi＝Pi′(xi)，hi＝xi+1-xi；
S3, solving the tri-diagonal matrix equation set constructed in the step S2 by adopting a catch-up method to obtain diagonal parameters;
s4, reconstructing a cubic equation among the points according to the diagonal parameters obtained in the step S3 to obtain a cubic spline difference coefficient, which is specifically expressed as:
s5, in each interval xi < x < xi +1, drawing out each equation reconstructed in the step S4 point by point to obtain a fitted spinal curve.
It will be appreciated by those of ordinary skill in the art that the embodiments described herein are intended to assist the reader in understanding the principles of the invention and are to be construed as being without limitation to such specifically recited embodiments and examples. Those skilled in the art can make various other specific changes and combinations based on the teachings of the present invention without departing from the spirit of the invention, and these changes and combinations are within the scope of the invention.
Claims (5)
1. A human back spine midline recognition method based on a depth camera is characterized by comprising the following steps:
A. acquiring a depth image of the back of a human body by using a depth camera;
B. b, gridding the depth image of the back of the human body obtained in the step A to obtain a grid matrix;
C. b, carrying out curvature processing on the grid matrix obtained in the step B to obtain a curvature matrix; the method specifically comprises the following steps: b, calculating a second derivative and a first derivative of the grid matrix obtained in the step B along the horizontal direction, and then combining the obtained second derivative and the obtained first derivative by adopting a curvature formula to obtain a curvature matrix; the curvature formula is expressed as:
wherein k represents curvature, and y is a depth image obtained by a depth camera;
setting the curvature value with negative curvature in the curvature matrix to zero, setting the curvature value with curvature larger than 1 in the curvature matrix to zero, and setting the isolated curvature value in the curvature matrix to zero;
D. c, extracting a central line region of the back according to the curvature matrix obtained in the step C; the method specifically comprises the following steps: carrying out corrosion operation on the curvature matrix, and dividing the curvature matrix into a plurality of characteristic areas; then removing isolated points and carrying out expansion operation; extracting the maximum connected domain of the curvature matrix to obtain a back midline region;
E. obtaining the coordinates of the keel points and the coordinates of the crotch points, and performing cubic spline interpolation fitting processing by combining the back central line region obtained in the step D to obtain the spine central line of the back of the human body; the method specifically comprises the following steps: obtaining a keel point coordinate and a crotch point coordinate and converting the keel point coordinate and the crotch point coordinate into a depth image coordinate; d, taking a median boundary value from the left boundary and the right boundary of the dorsal midline region obtained in the step D to obtain a partial spinal midline; and selecting the coordinates of the discrete points on the straight line connecting the coordinates of the keel points and the upper vertexes of the central lines of the partial spines and the coordinates of the discrete points on the straight line connecting the coordinates of the middle points of the connecting lines of the two crotch bone points and the lower vertexes of the central lines of the partial spines, and performing cubic spline interpolation fitting processing by combining the central lines of the partial spines to obtain the central line of the spine of the human back.
2. The method for recognizing the spine midline of the human back based on the depth camera as claimed in claim 1, wherein the selecting of the discrete point coordinates on the straight line connecting the carina point coordinates with the upper vertex of the part of the spine and the discrete point coordinates on the straight line connecting the midpoint coordinates of the two crotch point connecting lines with the lower vertex of the part of the spine, and the performing of cubic spline interpolation fitting processing in combination with the part of the spine midline to obtain the spine midline of the human back specifically comprises the following substeps:
s1, constructing a cubic spline interpolation fitting model;
s2, determining the boundary coordinates of the equation according to the connecting line midpoint between the keel point and the two crotch points, and constructing a tri-diagonal matrix equation set;
s3, solving the tri-diagonal matrix equation set constructed in the step S2 by adopting a catch-up method to obtain diagonal parameters;
s4, reconstructing a cubic equation among the points according to the diagonal parameters obtained in the step S3 to obtain a cubic spline difference coefficient;
s5, in each interval xi < x < xi +1, drawing out each equation reconstructed in the step S4 point by point to obtain a fitted spinal curve.
3. The depth camera-based human back spine midline recognition method of claim 2, wherein the cubic spline interpolation fitting model is specifically represented as:
Pi(x)＝ai+bi(x-xi)+ci(x-xi)2+di(x-xi)3
wherein, ai,bi,ci,diIs a cubic spline coefficient of difference, xiIs the x coordinate of the i-th discrete point, x being the generated fitThe x-coordinate of the curve.
4. The depth camera-based human back spinal midline recognition method of claim 3, wherein the tri-diagonal matrix equation set is specifically represented as:
wherein m isiDerivative of the fitted curve for the i-th segment, hiIs the y coordinate of the i-th discrete point, mi＝Pi′(xi)，hi＝xi+1-xi。
5. The depth camera-based human back spine midline recognition method of claim 4, wherein the cubic spline difference coefficient is specifically represented as:
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810479290.0A CN108764089B (en) | 2018-05-18 | 2018-05-18 | Human back spine midline recognition method based on depth camera |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810479290.0A CN108764089B (en) | 2018-05-18 | 2018-05-18 | Human back spine midline recognition method based on depth camera |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108764089A CN108764089A (en) | 2018-11-06 |
| CN108764089B true CN108764089B (en) | 2021-04-13 |
Family
ID=64007206
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810479290.0A Active CN108764089B (en) | 2018-05-18 | 2018-05-18 | Human back spine midline recognition method based on depth camera |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108764089B (en) |
Families Citing this family (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109919022A (en) * | 2019-01-29 | 2019-06-21 | 浙江工业大学 | A kind of adaptive inside and outside OCT fingerprint extraction method |
| CN109903277A (en) * | 2019-02-25 | 2019-06-18 | 电子科技大学 | A kind of scoliosis detection method based on polynomial curve fitting |
| CN111685770B (en) * | 2019-03-12 | 2023-03-10 | 香港理工大学深圳研究院 | Wearable human body back curve detection method and device |
| CN111598618B (en) * | 2020-05-14 | 2023-06-23 | 武汉轻工大学 | Three-dimensional curved surface data estimation method, device, equipment and storage medium |
| CN114224322B (en) * | 2021-10-25 | 2023-06-02 | 上海工程技术大学 | Scoliosis assessment method based on key points of human bones |
| CN113951874B (en) * | 2021-10-25 | 2023-12-26 | 中国科学院长春光学精密机械与物理研究所 | Scoliosis risk assessment system |
| CN115272588B (en) * | 2022-09-27 | 2022-12-30 | 广州辉博信息技术有限公司 | Human body spine fitting method and system based on depth image and electronic equipment |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106875377A (en) * | 2016-12-30 | 2017-06-20 | 西北工业大学 | A kind of vertebra characteristic point automatic identifying method based on Gaussian curvature stream |
| CN107464275A (en) * | 2017-07-28 | 2017-12-12 | 电子科技大学 | Human spine center line three-dimensional reconstruction method |
| CN107481228A (en) * | 2017-07-28 | 2017-12-15 | 电子科技大学 | Human body back scoliosis angle measurement method based on computer vision |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20060036250A1 (en) * | 2004-08-12 | 2006-02-16 | Lange Eric C | Antero-lateral plating systems for spinal stabilization |
-
2018
- 2018-05-18 CN CN201810479290.0A patent/CN108764089B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106875377A (en) * | 2016-12-30 | 2017-06-20 | 西北工业大学 | A kind of vertebra characteristic point automatic identifying method based on Gaussian curvature stream |
| CN107464275A (en) * | 2017-07-28 | 2017-12-12 | 电子科技大学 | Human spine center line three-dimensional reconstruction method |
| CN107481228A (en) * | 2017-07-28 | 2017-12-15 | 电子科技大学 | Human body back scoliosis angle measurement method based on computer vision |
Also Published As
| Publication number | Publication date |
|---|---|
| CN108764089A (en) | 2018-11-06 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108764089B (en) | Human back spine midline recognition method based on depth camera | |
| CN109409437B (en) | Point cloud segmentation method and device, computer readable storage medium and terminal | |
| CN103913131B (en) | Free curve method vector measurement method based on binocular vision | |
| Lee et al. | Curved glide-reflection symmetry detection | |
| CN112200843B (en) | Super-voxel-based CBCT and laser scanning point cloud data tooth registration method | |
| CN102722890B (en) | Non-rigid heart image grading and registering method based on optical flow field model | |
| CN105160686B (en) | A kind of low latitude various visual angles Remote Sensing Images Matching Method based on improvement SIFT operators | |
| CN109903277A (en) | A kind of scoliosis detection method based on polynomial curve fitting | |
| CN104268515A (en) | Sperm morphology anomaly detection method | |
| CN104809457A (en) | Three-dimensional face identification method and system based on regionalization implicit function features | |
| CN109727260A (en) | A kind of three-dimensional lobe of the lung dividing method based on CT images | |
| CN107240128B (en) | X-ray and color photo registration method based on contour features | |
| CN109886124A (en) | One kind describing the matched texture-free metal parts grasping means of subgraph based on harness | |
| CN112330813A (en) | Wearing three-dimensional human body model reconstruction method based on monocular depth camera | |
| CN112085675A (en) | Depth image denoising method, foreground segmentation method and human motion monitoring method | |
| CN111862315A (en) | Human body multi-size measuring method and system based on depth camera | |
| CN108280833B (en) | Skeleton extraction method for plant root system bifurcation characteristics | |
| CN109241822A (en) | A kind of multi-faceted method for detecting human face and system based on MTCNN | |
| CN110223331B (en) | Brain MR medical image registration method | |
| CN113706514B (en) | Focus positioning method, device, equipment and storage medium based on template image | |
| CN110047146B (en) | Error correction method based on single revolving body image 3D restoration | |
| CN105488798B (en) | SAR image method for measuring similarity based on point set contrast | |
| CN104899923A (en) | Method for constructing facial prosthesis optimized model based on smile expression geometrical characteristic modification | |
| CN108447038A (en) | A kind of mesh denoising method based on non local full variation operator | |
| CN109509194B (en) | Front human body image segmentation method and device under complex background |
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 | ||
| TR01 | Transfer of patent right | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20211108 Address after: 518057 314, industry university research building, Hong Kong Polytechnic University, 18 Yuexing 1st Road, high tech Zone community, Yuehai street, Nanshan District, Shenzhen, Guangdong Patentee after: Shenzhen Lingdong Medical Technology Co.,Ltd. Address before: Chengdu, Sichuan Province Patentee before: SICHUAN EFFICIENCY FUTURE TECHNOLOGY Co.,Ltd. |