CN102831443A - Skull sex determining method based on spatial analysis - Google Patents
Skull sex determining method based on spatial analysis Download PDFInfo
- Publication number
- CN102831443A CN102831443A CN2012102646300A CN201210264630A CN102831443A CN 102831443 A CN102831443 A CN 102831443A CN 2012102646300 A CN2012102646300 A CN 2012102646300A CN 201210264630 A CN201210264630 A CN 201210264630A CN 102831443 A CN102831443 A CN 102831443A
- Authority
- CN
- China
- Prior art keywords
- skull
- data
- subspace
- sex
- training sample
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Images
Landscapes
- Image Processing (AREA)
- Analysing Materials By The Use Of Radiation (AREA)
Abstract
The invention relates to a skull sex determining method based on spatial analysis. The method comprises the following steps: carrying out data standardization on skull data of a training sample; carrying out sub-spatial analysis on a standardized training sample data set so as to obtain a sub-spatial projection matrix; projecting the training sample to a low dimensional sub-space from a high dimensional original feature space, wherein a classifier is designed in the low dimensional sub-space; projecting standardized unknown skull data to the low dimensional sub-space by using a sub-spatial projection matrix obtained from the training; and classifying in the low dimensional sub-space by using the designed classifier. With the adoption of the method, complex manual measurement or expert knowledge is not needed, and the accuracy can be more than 94%, so that the method has important application values in criminal investigation fields.
Description
Technical field
The present invention relates to the Computer Applied Technology field, particularly a kind of method of utilizing the subspace characteristic to carry out the classification of skull sex.Be mainly used in fields such as criminal investigation, archaeology, forensic anthropology.
Background technology
Carrying out sex identification according to the skeleton form is one of research contents the most basic in the forensic anthropology.Skull is to be made up of sclerous tissues, and other bone is difficult for being destroyed relatively, after death also can more intactly retain down.Under many circumstances, detecting site is only retained victim's skull, does not have other body source clue, and this makes that classical inspection technologies such as DNA, photographic superimposition are lack scope for their abilities, only can rely on skull to carry out the corpse source and identify, and identify that sex is a first step wherein.With other primate particularly orangutan compare with baboon, the gender differences that human skull embodied are very little, and can receive the influence of heredity, environment, nutrition condition and behavioral activity, these have all increased the difficulty of skull sex identification.Identify that according to skull sex has become the research focus of association areas such as information science, anthropology, medical jurisprudence in the world.
Existent method is divided into two types at present: form diagnostic method and measurement diagnostic method.The form diagnostic method mainly is to be identified according to its understanding to skull morphological feature gender differences by the anthropology expert, and is more obvious such as male sex's geisoma and nasal bone, mandibular and frontal bone side, and women's frontal eminence is outstanding or the like than the male sex.The form diagnostic method seriously relies on the subjectivity that the expert understands morphological feature, and different experts' understanding has certain difference.Measure diagnostic method manual some measurement points of demarcating on the plaster statue of the X-ray photograph of skull or skull usually, utilize measurement point to set up some how much variablees, according to these how much variablees, through statistics structure discriminant function to some samples.Than the form diagnostic method, measure diagnostic method to expertise require relatively low, but the measuring accuracy of geometric sense is required very high, for the biosome of complicacy like this, realize the accurate very difficulty of measuring.Existing document shows that as far as most of characteristic, the measuring error between the different observers reaches more than 10%.In addition, after growing up, with advancing age, the form of skull can not change basically, but size can change.These have all increased measures the difficulty of differentiating.Through digitizing, measure diagnostic method and can realize by area of computer aided to skull.Fordisc is the computer assisted bone analysis software that American Jantz and Ousley developed in 1993, has been upgraded to the third edition at present.This software adopts the skull data in U.S.'s criminal investigation database to set up the sex discriminant function through some geometric senses of interactive means measurement bone.But to realize also not a duck soup of accurate pattern measurement on computers.In addition, present method is not high to the correct identification rate of skull test sample book collection, is no more than 90% basically.
Summary of the invention
The object of the invention overcomes deficiency of the prior art exactly, and the skull sex appraisal method based on subspace analysis that a kind of automaticity is high, applied widely, accuracy is higher is provided.
For realizing above-mentioned purpose, technical solution of the present invention is to utilize the subspace analysis technology that the skull data projection of higher-dimension is arrived low subspace of tieing up, and extracts the subspace characteristic, at the low subspace design category device of tieing up.
The key step of method of the present invention comprises:
1) classification based training
1.1) data requirementization: the skull data to training sample is concentrated are carried out non-rigid data registration, with the attitude and the size specificationization of skull data, at last each skull data are arranged as the original feature vector of a higher-dimension by the coordinate of point;
1.2) subspace analysis: normalized training sample data collection is carried out subspace analysis, obtain the subspace projection matrix;
1.3) classifier design: with training sample from higher-dimension primitive character space projection to low n-dimensional subspace n, at low n-dimensional subspace n design category device;
2) sex identification
2.1) data requirementization: the skull data that unknown skull and training sample are concentrated are carried out non-rigid data registration,, and be arranged as a higher-dimension original feature vector by the coordinate of putting with its attitude and size specificationization;
2.2) the subspace feature extraction: utilize step 1.2) in the subspace projection matrix with the standardizing number of unknown skull according to projecting to low n-dimensional subspace n;
2.3) sex classification: utilize step 1.3) in sorter carry out sex identification.
Preferably, skull data step 1.1) are X-ray photograph data or 3D grid data of two dimension.
Preferably, subspace analysis step 1.2) adopts independent component to analyze a kind of among ICA, the principal component analysis PCA.
Preferably, the classifier design step 1.3) adopts Fisher linear discriminant analysis method.
The present invention is based on the skull sex appraisal method of subspace analysis, the skull data dimensionality reduction of higher-dimension to low n-dimensional subspace n, is carried out sex at low n-dimensional subspace n and classifies.The characteristics of this method are: skull has been carried out big or small normalization, has removed the influence of skull size, utilization be the shape information of skull; Need not fussy hand and measure, need not expertise, automaticity is high, and accuracy is higher, reaches more than 94%.
Description of drawings
Fig. 1 is the process flow diagram that the present invention is based on the skull sex appraisal method of subspace analysis.
Embodiment
In order more clearly to introduce technical scheme of the present invention, the present invention is done further detailed explanation below in conjunction with accompanying drawing.
As shown in Figure 1, the key step of the skull sex appraisal method based on subspace analysis according to the invention comprises:
1) classification based training
1.1) data requirementization: the skull data to training sample is concentrated are carried out non-rigid data registration, with the attitude and the size specificationization of skull data, at last each skull data are arranged as the original feature vector of a higher-dimension by the coordinate of point;
1.2) subspace analysis: normalized training sample data collection is carried out subspace analysis, obtain the subspace projection matrix;
1.3) classifier design: with training sample from higher-dimension primitive character space projection to low n-dimensional subspace n, at low n-dimensional subspace n design category device;
2) sex identification
2.1) data requirementization: the skull data that unknown skull and training sample are concentrated are carried out non-rigid data registration,, and be arranged as a higher-dimension original feature vector by the coordinate of putting with its attitude and size specificationization;
2.2) the subspace feature extraction: utilize step 1.2) in the subspace projection matrix with the standardizing number of unknown skull according to projecting to low n-dimensional subspace n;
2.3) sex classification: utilize step 1.3) in sorter carry out sex identification.
Described skull data among the present invention are the X-ray photograph data or the 3D grid data of two dimension; The difference that both handle only is step 1.1) non-rigid data registration in the data requirementization; But the X-ray photograph data or the non-rigid data registration of 3D grid data that are two dimension all are the more problems of research, can both adopt corresponding classical way.Describe for the 3D grid data to described skull data below.The three-dimensional cranium data that adopted are through the Cranial Computed Tomography image data is carried out cutting apart of skull and musculus cutaneus; And the 3D grid data that adopt the reconstruct of Marching Cubes method to obtain; Each 3D grid data comprises about 150,000 summits and 300,000 tri patchs.Supposing has n training sample, below is concrete implementation step:
The step 1) classification based training comprises as follows step by step:
Step 1.1); Data requirementization: arrive Frankfort coordinate system to the 3D grid uniform data of the concentrated skull of training sample; With the distance of the left and right earhole central point of each sample distance metric unit, promptly the coordinate of being had a few on the sample is carried out yardstick normalization as this sample.Select one with reference to skull, the skull data that training sample is concentrated are carried out three-dimensional non-rigid data registration, obtain the grid model that number of vertices is identical, semanteme is identical, annexation is consistent.Three-dimensional non-rigid data registration is one has a large amount of methods to supply to select for use than proven technique, and TPS (the Thin Plate Spline) method for registering with a kind of iteration is an example here, the steps include:
1. with reference to skull S
0On select some (N > at random; 6) some M
0={ L
0j| L
0j=(x
0j, y
0j, z
0j), j=1 ..., N} is as unique point, and on target skull T, obtains its corresponding point M through ICP (Iterative Closest Point) method
2={ L
2j| L
2j=(x
2j, y
2j, z
2j), j=1 ..., N} to the control point set of corresponding point as the TPS conversion, asks for S with this N
0And the TPS conversion f between T, to S
0Carry out obtaining behind the conversion f after the conversion with reference to skull S
1
2. with reference to skull S
0On select some (N > again at random; 6) point
M
0={ L
0j| L
0j=(x
0j, y
0j, z
0j), j=1 ..., N}, the TPS conversion f that 1. obtains according to step asks for point set M
0At S
1On corresponding point set M
1={ L
1j| L
1j=(x
1j, y
1j, z
1j), j=1 ..., N} passes through S again
1T carries out ICP with the target skull, obtains M
1Corresponding point set M on target skull T
2={ L
2j| L
2j=(x
2j, y
2j, z
2j), j=1 ..., N} is with M
0, M
2These group corresponding point are asked for S again as new control point set
0And the TPS conversion f between T, the reference skull S after the renewal conversion
1
If 3. S
1And the error sum of corresponding point reaches given threshold value between target skull T, or number of times that surpass to set of iterations, changes next step over to, otherwise changes step 2.;
4. according to the last TPS conversion f that obtains confirm target skull T with reference to skull S
0Between point corresponding, utilize a corresponding rotation and the translation transformation that calculates between the two of point, and the target skull carried out conversion to realize the attitude adjustment.
Threshold value and the iterations of step in 3. can rule of thumb be set.
Suppose that each the skull data behind the registration comprise m summit { (x
iy
iz
i), i=1,2 ..., m} is arranged as each skull data the original feature vector (x of a 3m dimension by the coordinate on summit
1, y
1, z
1, x
2, y
2, z
2..., x
m, y
m, z
m)
T
Step 1.2); Subspace analysis: normalized training sample data collection is carried out subspace analysis, and subspace analysis adopts independent component to analyze ICA, principal component analysis PCA etc., is example to adopt principal component analysis PCA; Keep 95% data variance, obtain a projection matrix P
3m * k, k wherein<n.
Step 1.3), classifier design: utilize projection matrix that training sample is tieed up the primitive character space projection to the k n-dimensional subspace n from 3m, each training sample obtains a k n-dimensional subspace n characteristic, utilizes Fisher linear discriminant analysis design category device at the k n-dimensional subspace n.
Step 2), sex identification
Step 2.1); Data requirementization: the 3D grid uniform data of unknown skull to be identified to Frankfort coordinate system and carry out yardstick normalization; Carry out as step 1.1) in three-dimensional non-rigid data registration, the 3D grid data of the unknown skull behind the registration are arranged as the original feature vector of a 3m dimension by the coordinate on summit.
Step 2.2), the projection matrix subspace feature extraction: utilize step 1.2) projects to the k n-dimensional subspace n with the standardizing number certificate of target skull.
Step 2.3), the sorter sex classification: utilize step 1.3) carries out sex identification.
In a word, what embodiments of the invention were announced is its preferred implementation, but is not limited to this.Those skilled in the art understands spirit of the present invention very easily according to the foregoing description, only otherwise the modification or the replacement that break away from the basis of technical scheme of the present invention, and all within protection scope of the present invention.
Claims (4)
1. based on the skull sex appraisal method of subspace analysis, it is characterized in that, mainly may further comprise the steps:
1) classification based training
1.1) data requirementization: the skull data to training sample is concentrated are carried out non-rigid data registration, with the attitude and the size specificationization of skull data, at last each skull data are arranged as the original feature vector of a higher-dimension by the coordinate of point;
1.2) subspace analysis: normalized training sample data collection is carried out subspace analysis, obtain the subspace projection matrix;
1.3) classifier design: with training sample from higher-dimension primitive character space projection to low n-dimensional subspace n, at low n-dimensional subspace n design category device;
2) sex identification
2.1) data requirementization: the skull data that unknown skull and training sample are concentrated are carried out non-rigid data registration,, and be arranged as a higher-dimension original feature vector by the coordinate of putting with its attitude and size specificationization;
2.2) the subspace feature extraction: utilize step 1.2) in the subspace projection matrix with the standardizing number of unknown skull according to projecting to low n-dimensional subspace n;
2.3) sex classification: utilize step 1.3) in sorter carry out sex identification.
2. skull sex appraisal method according to claim 1 is characterized in that step 1.1) described in the skull data be the two dimension X-ray photograph data or 3D grid data.
3. skull sex appraisal method according to claim 1 and 2 is characterized in that step 1.2) described in subspace analysis adopt independent component to analyze a kind of among ICA, the principal component analysis PCA.
4. skull sex appraisal method according to claim 1 and 2 is characterized in that step 1.3) described in classifier design adopt Fisher linear discriminant analysis method.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201210264630.0A CN102831443B (en) | 2012-07-27 | 2012-07-27 | Skull sex determining method based on spatial analysis |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201210264630.0A CN102831443B (en) | 2012-07-27 | 2012-07-27 | Skull sex determining method based on spatial analysis |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN102831443A true CN102831443A (en) | 2012-12-19 |
| CN102831443B CN102831443B (en) | 2015-03-18 |
Family
ID=47334569
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201210264630.0A Active CN102831443B (en) | 2012-07-27 | 2012-07-27 | Skull sex determining method based on spatial analysis |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN102831443B (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104537642A (en) * | 2014-12-08 | 2015-04-22 | 电子科技大学 | Malignant glioblastoma multiforme tissue classification method based on non-negative matrix factorization |
| CN104851123A (en) * | 2014-02-13 | 2015-08-19 | 北京师范大学 | Three-dimensional human face change simulation method |
| CN108197539A (en) * | 2017-12-21 | 2018-06-22 | 西北大学 | A kind of Diagnosis of Crania By Means identification method |
| CN108647733A (en) * | 2018-05-15 | 2018-10-12 | 西北大学 | A kind of breakage Diagnosis of Crania By Means identification method |
| CN112907537A (en) * | 2021-02-20 | 2021-06-04 | 司法鉴定科学研究院 | Skeleton sex identification method based on deep learning and on-site virtual simulation technology |
-
2012
- 2012-07-27 CN CN201210264630.0A patent/CN102831443B/en active Active
Non-Patent Citations (3)
| Title |
|---|
| DIRK VANDERMEULEN EL AT.: "《Computerized craniofacial reconstruction using CT-derived implicit surface representations》", 《ELSEVIER:FORENSIC SCIENCE INTERNATIONAL》 * |
| 朱新懿等: "《利用改进的分区统计颅面模型重构颅面》", 《计算机工程与引用》 * |
| 武勃等: "《人脸的性别分类》", 《计算机研究与发展》 * |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104851123A (en) * | 2014-02-13 | 2015-08-19 | 北京师范大学 | Three-dimensional human face change simulation method |
| CN104851123B (en) * | 2014-02-13 | 2018-02-06 | 北京师范大学 | A kind of three-dimensional face change modeling method |
| CN104537642A (en) * | 2014-12-08 | 2015-04-22 | 电子科技大学 | Malignant glioblastoma multiforme tissue classification method based on non-negative matrix factorization |
| CN104537642B (en) * | 2014-12-08 | 2018-09-04 | 电子科技大学 | A kind of malignant glioblastoma tissue classification procedure based on Non-negative Matrix Factorization |
| CN108197539A (en) * | 2017-12-21 | 2018-06-22 | 西北大学 | A kind of Diagnosis of Crania By Means identification method |
| CN108647733A (en) * | 2018-05-15 | 2018-10-12 | 西北大学 | A kind of breakage Diagnosis of Crania By Means identification method |
| CN112907537A (en) * | 2021-02-20 | 2021-06-04 | 司法鉴定科学研究院 | Skeleton sex identification method based on deep learning and on-site virtual simulation technology |
Also Published As
| Publication number | Publication date |
|---|---|
| CN102831443B (en) | 2015-03-18 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Tsagkrasoulis et al. | Heritability maps of human face morphology through large-scale automated three-dimensional phenotyping | |
| Hill et al. | Model-based interpretation of 3d medical images. | |
| Duan et al. | 3D face reconstruction from skull by regression modeling in shape parameter spaces | |
| Shui et al. | A PCA-Based method for determining craniofacial relationship and sexual dimorphism of facial shapes | |
| Devine et al. | A registration and deep learning approach to automated landmark detection for geometric morphometrics | |
| Li et al. | Rapid automated landmarking for morphometric analysis of three‐dimensional facial scans | |
| CN102831443A (en) | Skull sex determining method based on spatial analysis | |
| Deng et al. | A regional method for craniofacial reconstruction based on coordinate adjustments and a new fusion strategy | |
| Bano et al. | AutoFB: Automating fetal biometry estimation from standard ultrasound planes | |
| Abu et al. | Automated craniofacial landmarks detection on 3D image using geometry characteristics information | |
| Guo et al. | Cross-dataset evaluation of deep learning networks for uterine cervix segmentation | |
| Vandermeulen et al. | Automated facial reconstruction | |
| Thual et al. | Aligning individual brains with fused unbalanced Gromov Wasserstein | |
| Young et al. | Performance of single and multi-atlas based automated landmarking methods compared to expert annotations in volumetric microCT datasets of mouse mandibles | |
| Manthey et al. | Geometric morphometrics | |
| Maranho et al. | Sexual dimorphism of the human scapula: a geometric morphometrics study in two Portuguese reference skeletal samples | |
| Hu et al. | Craniofacial reconstruction based on a hierarchical dense deformable model | |
| Wilson et al. | Voyaging into the third dimension: A perspective on virtual methods and their application to studies of juvenile sex estimation and the ontogeny of sexual dimorphism | |
| Avisdris et al. | Biometrynet: landmark-based fetal biometry estimation from standard ultrasound planes | |
| Xue et al. | Multimodal patient-specific registration for breast imaging using biomechanical modeling with reference to AI evaluation of breast tumor change | |
| CN103310440A (en) | Canonical correlation analysis-based skull identity authentication method | |
| Farnell et al. | What’s in a Smile? Initial results of multilevel principal components analysis of facial shape and image texture | |
| Rahim et al. | A diffeomorphic mapping based characterization of temporal sequences: application to the pelvic organ dynamics assessment | |
| Fontanella et al. | The offset normal shape distribution for dynamic shape analysis | |
| Ebsim et al. | Fully automatic detection of distal radius fractures from posteroanterior and lateral radiographs |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant |