CN100462054C - Method for shielding sex part on foetus image for preventing recognizing foetus sex - Google Patents
Method for shielding sex part on foetus image for preventing recognizing foetus sex Download PDFInfo
- Publication number
- CN100462054C CN100462054C CNB2007100759017A CN200710075901A CN100462054C CN 100462054 C CN100462054 C CN 100462054C CN B2007100759017 A CNB2007100759017 A CN B2007100759017A CN 200710075901 A CN200710075901 A CN 200710075901A CN 100462054 C CN100462054 C CN 100462054C
- Authority
- CN
- China
- Prior art keywords
- sex
- image
- foetus
- primary location
- location point
- 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
Landscapes
- Image Processing (AREA)
- Ultra Sonic Daignosis Equipment (AREA)
Abstract
The present invention is method of recognizing and shielding the sex part in fetus image includes the following steps: 1. processing the original fetus image, finding out the edge contour of the sex part, and determining at least one initial locating point of suspected sex part; 2. recognizing the initial locating point by means of the fetus sex image database and recognizing system with multiple classifier fusion structure to judge the fetus sex part; and 3. shielding the fetus sex part with shielding image. The present invention makes it possible to avoid illegal fetus sex identification in ultrasonic apparatus effectively.
Description
Technical field
The present invention relates to a kind of method that fetus image sex position is carried out Intelligent Recognition and shielded with shielding patterns.
Background technology
The application of ultrasonic device is increasingly extensive.For example, be applied in the medical diagnosis of human body internal organs and detection fetal growth situation.Ultrasonic device has been made major contribution for human beings'health on the one hand, also produces some negative effects on the other hand.China's population has surpassed 1,300,000,000; huge population pressure has brought white elephant for many aspects such as China's economic development, social stability and environmental conservation; therefore; family planning has been made huge contribution to slowing down China's rate of population increase with the pressure that the alleviation overpopulation brings as the fundamental state policy of China.Yet China rural area is subjected to that "males are better than females", and this falls behind the influence of custom, uses ultrasonic device to carry out that illegal sex of foetus is identified and the people is still very serious as the interruption of pregnancy phenomenon selectively.China's ewborn infant sex ratio that national the 5th census in 2000 is announced is up to 116:100, considerably beyond the international the highest warning line 107:100 that can tolerate.The newborn population sex ratio is serious unbalance certainly will to constitute threat greatly to China's economic development, social stability.
In view of this, be necessary to research and develop and a kind ofly can effectively prevent to utilize the ultrasonoscopy selectivity to carry out the method that the people is taken place for the interruption of pregnancy phenomenon.
Summary of the invention
The purpose of this invention is to provide a kind of method for shielding sex part on foetus image for preventing recognizing foetus sex, this method can be discerned and shield with shielding patterns the sex of foetus position in the original image effectively.
For achieving the above object, the method for shielding sex part on foetus image for preventing recognizing foetus sex of the present invention's proposition comprises the following steps:
A. the original image that generates is carried out Flame Image Process, find out, determine the Primary Location point at least one doubtful shielding position by intending the edge wheel profile at shielding position;
B. use the recognition system of sex of foetus view data parameter library and multiple Classifiers Combination structure that the Primary Location point at doubtful shielding position is discerned, judge and determine the shielding position;
C. use shielded image to shield to the shielding position of determining.
Compared with prior art, the method for shielding sex part on foetus image for preventing recognizing foetus sex that the present invention proposes can shield the sex of foetus position in the image easily, and the information at other positions and original image are identical, replace original image to play demonstration the image after handling, the doctor can't correctly judge sex of foetus from image, thereby reach illegal sex of foetus is identified the purpose that phenomenon is contained.
Description of drawings
Fig. 1 is the theory diagram of method for shielding sex part on foetus image for preventing recognizing foetus sex of the present invention;
Fig. 2 is a particular flow sheet of fetus image sex position being discerned screen method of the present invention;
Fig. 3 is the flow chart of adaptively selected segmentation threshold method among the present invention.
The specific embodiment
Key problem in technology of the present invention is the fetus image is carried out Intelligent Recognition, and whether the image that is automatically identified when pre-treatment by machine contains sex of foetus position and the sex of foetus position particular location in image.Know-why of the present invention comprises Flame Image Process, pattern recognition and artificial intelligence technology.
Fig. 1 is the theory diagram that the present invention is applied in a better embodiment in the ultrasonography.General ultrasonic device 10 comprises a ultrasonoscopy sniffer 11, a ultrasonoscopy generating apparatus 12 and an image display device 14.Ultrasonoscopy sniffer 11 adopts ultrasound wave that the fetus image in the pregnant woman uterus is surveyed; The signal that ultrasonoscopy generating apparatus 12 is used for detecting according to ultrasonoscopy sniffer 11 generates the fetus image; 14 of image display devices are used to show the fetus image.The present invention also is provided with an image processing apparatus 13 between video generation device 12 and image display device 14, the sex position that is used for fetus image that ultrasonoscopy generating apparatus 12 is generated is discerned and shielded.
Sex of foetus view data parameter library 15 among the present invention obtains from the image library that 20000 width of cloth typical case fetus image is formed according to the concrete supervised learning grader that adopts, 9273 width of cloth image packets contain the sex of foetus position in this image library, and 10727 width of cloth images do not comprise the sex of foetus position.The fetus image of these gestational ages from 16 weeks to 34 weeks do not limit anemia of pregnant woman's position from the ultrasonic examination collection in worksite during collection, comprise multiple positions such as dorsal position, right arm reclining and left lateral position.The particular location that whether comprises sex of foetus position and sex position in the image is confirmed jointly by several veteran gynecologists.This image pattern storehouse has fully covered the sex of foetus station diagram picture that reaches different gestational age under different anemia of pregnant woman's positions, is a standard picture storehouse describing sex of foetus station diagram picture.
Fig. 2 discerns the particular flow sheet of screen method to the sex of foetus position for the present invention.The key step of the inventive method is as follows:
S10: the original ultrasonoscopy to input carries out the image pretreatment, and this step has this image pretreatment module 16 to finish.
Signal to noise ratio is low, the speckle noise phenomenon serious because the restriction of ultrasonic device image-forming principle, ultrasonoscopy exist and characteristics such as imaging object edge blurry, therefore need carry out pretreatment to original ultrasonoscopy.Can adopt auto-adaptive filtering technique that original ultrasonoscopy is carried out pretreatment, this technology is at first analyzed the whole signal noise ratio level and the speckle noise order of severity of ultrasonoscopy, then each picture element in the image is all judged that by the intensity profile situation of local neighborhood around it this pixel is positioned at homogeneous region or different filtering modes is selected in the non-homogeneous zone.Under the occasion that algorithm speed is had relatively high expectations, can directly use traditional filtering modes such as mean filter that image is carried out pretreatment.
S20: pretreated ultrasonoscopy is carried out image segmentation, and this step is finished by image segmentation module 17.
Because ultrasonic fetus image imaging content is very complicated, and the sex position imaging difference of different all numbers, different anemia of pregnant woman's position and different fetus individualities is very big, if conventional art means such as pretreated ultrasonoscopy use template similarity coupling are unpractical and do not have feasibility.In order further to extract the information in the image, need carry out image segmentation to pretreated image and handle.Dividing method for ultrasonoscopy can be divided into two classes according to the situation of practical application platform.The occasion that the first kind is had relatively high expectations at algorithm speed is used the fixed threshold classification method, and segmentation threshold can select 80 or 60 according to the integral image grey level.Second class is used the adaptive threshold system of selection based on the supervised learning technology for obtaining more accurate classifying quality.
Fig. 3 is the flow chart of adaptively selected segmentation threshold method among the present invention.At first with parts of images in the fetus sample image storehouse as training image, extract overall intensity features such as gray average and variance as characteristic information, in several pre-selected threshold, select only segmentation threshold as label (Label) information again, form training sample jointly.Use BP neutral net or support vector machine supervised learning graders such as (SVM) that training sample is trained, with another part image in the fetus sample image storehouse as test pattern, extract test sample book and test, thereby obtain the best classifier parameters of effect.The parameter that obtains is deposited into sex of foetus view data parameter library 15.When using, use the relevant parameter of depositing in the sex of foetus view data parameter library 15 that grader is carried out parameter configuration, new need split image is extracted the overall intensity feature, obtain segmentation threshold automatically by the grader that configures parameter again, carry out image segmentation.
S30: extract the Primary Location point at doubtful sex of foetus position, this step is finished by Primary Location point extraction module 18.
The sex of foetus position only accounts for very little a part of area usually in the fetal ultrasound image, for example a width of cloth digital ultrasound image is made up of 768 * 576 picture elements, wherein the sex of foetus position generally accounts for 50 * 50 pixel sizes, and proportion only has about 5/1000ths in entire image.In order in image, accurately to locate the sex of fetus position, at first need coarse positioning is carried out at the sex position.According to sex of foetus position imaging characteristics, image object edge contour after cutting apart will pass through the sex of foetus position usually, and the contour line at fetus position has the high feature of degree of crook, therefore need to extract the edge wheel profile of cutting apart the back object, and the contour line curvature value extracted, and obtain the curvature extreme point as sex of foetus position Primary Location point.
Contour of object line drawing after the image segmentation uses the profile tracking technique to finish, and uses chained list to represent by the continuous adjacent mode complete contour line.Curvature is described the degree of crook of line segment, has reflected the tendency information of line segment, and the line segment curvature value can use 11 methods or B batten method to extract.Because sex position contour line degree of crook height, pairing curvature absolute value is bigger, and the curvature value to each picture element on the contour line carries out analytical judgment subsequently, selects the bigger picture element of its mean curvature absolute value as anchor point.The scope that the Primary Location point has dwindled image recognition has improved the levels of precision of identification.
S40: judge whether doubtful sex position is positioned at the sex of foetus position, and this step is finished by sex of foetus part Identification module 19.
The sex position Primary Location number of spots of one width of cloth fetus image preliminary election has nothing in common with each other according to picture material is different, usually about 10 to 50, if include the sex of foetus position in the image, these Primary Location have been named a person for a particular job and one have been positioned at inside, sex of foetus position to several anchor points, and the first anchor point that therefore is positioned at non-sex position need further be excluded; If do not comprise the sex position in the image, then all first anchor points all need to be excluded.
Because ultrasonograph quality is relatively poor, in order to improve the accuracy rate of identification, this method is used the multiple Classifiers Combination framework whether each Primary Location point is positioned at the sex of foetus position and is discerned judgement.The grader principle all belongs to aforementioned supervised learning technology, concrete in the method grader can be the BP neural network classifier, what also can be main flow has a supervised classification device (as SVM).
In the present embodiment, the multiple Classifiers Combination framework is a double-layer structure, and ground floor is made up of a plurality of graders, and each grader is all accepted the different characteristic vector of Primary Location point and imported as grader, and grader is output as recognition credibility.The eigenvalue of Primary Location point is selected to comprise following several mode: 1, select continuous curvature value through the body outline of Primary Location point; 2, selecting with the Primary Location point is the spatial domain intensity profile in the topography zone at center, comprises information such as average, variance; 3, selecting with the Primary Location point is the picture element gray scale in the topography zone at center, and carries out the pyramid sample process.Different topography's area size and different sampling proportions have reflected with the Primary Location point to be the multi-level image feature of central area.The pyramid sampling approach is determined sampling proportion earlier, for example, sampling proportion is 4, illustrates that then the zonule that 4 * 4 picture elements are formed is sampled to 1 picture element, the method of sampling adopts averaging method in the present embodiment, promptly gets this zone meansigma methods of totally 16 picture element gray values.Be chosen as 100 * 100 as topography's area size, sampling proportion is chosen as 4, can obtain one group of totally 25 * 25=625 eigenvalue, forms a characteristic vector.According to the difference of topography's area size and sampling proportion, can obtain many stack features value.Based on the training of the supervised learning grader in fetus sample image storehouse and test process as hereinbefore, with the classifier parameters that obtains training.The parameter that obtains is deposited into sex of foetus view data parameter library 15.When using, use the relevant parameter of depositing in the sex of foetus view data parameter library 15 that grader is carried out parameter configuration, the Primary Location point of new need identification will obtain recognition credibility by the grader that has disposed parameter, be used as down the input of one deck grader.
The second layer is made up of a grader, that can adopt main flow has a supervised classification device, the weighting ballot method graders that adopt more, this grader is collected the credibility output of all graders of ground floor and is formed a credibility feature group, input feature vector as this grader, the credibility that grader is output as weighting after handling, output valve are between 0 to 1, and numerical value is big more, and to show that processed Primary Location point is positioned at the probability at sex position big more.The optimal weights parameter that weighting ballot method grader obtains in the training and testing process also will be deposited into sex of foetus view data parameter library 15.When using, use the respective weights parameter of depositing in the sex of foetus view data parameter library 15 that grader is configured, the credibility feature group of new need identification will obtain final weighting credibility by the grader that has disposed parameter, as the final result of identification.In the present embodiment, select output valve to represent the sex position, the non-sex of Primary Location point representative position smaller or equal to 0.8 greater than 0.8 Primary Location point.
S50: to the sex position of identification gained in addition mosaic or other shielding patterns shield, mosaic all is to be the center with the anchor point, its shape, size and pixel ash value also are adjustable.This step is finished by sex of foetus position shroud module 20.It is the center that the scope of adding mosaic in the present embodiment is chosen as with identification gained sex spots localization point, and 25 picture element sizes are the square area of radius, and the pixel gray value unification in this zone is revised as 128.The image pixel gray scale at other non-sex positions remains unchanged, thereby has obtained covering the ultrasonoscopy at sex of foetus position.
Compared to prior art, said method and equipment can be realized easily to the sex of foetus section in the ultrasonoscopy The position shields, and the information at other positions and original ultrasonoscopy are identical, and the image after processing is logical Cross hardware platform and send into ultrasonic device and replace original image to play demonstration, the doctor can't be from ultrasonoscopy Correct judgement sex of foetus, thus reach the purpose that illegal sex identification phenomenon is contained.
Claims (10)
1. a method for shielding sex part on foetus image for preventing recognizing foetus sex is characterized in that this method comprises the following steps:
A. the fetus original image that generates is carried out Flame Image Process, find out edge wheel profile, determine the Primary Location point at least one doubtful sex position by the sex of foetus position;
B. use the recognition system of sex of foetus view data parameter library and multiple Classifiers Combination structure that the Primary Location point at doubtful sex position is discerned, judge and determine the sex of foetus position;
C. use shielded image to shield to the sex of foetus position of determining.
2. the method for claim 1 is characterized in that: the Flame Image Process described in the step a comprises at first carries out image filtering denoising pretreatment to the ultrasonic fetus original image that generates.
3. method as claimed in claim 2, it is characterized in that: among the step a, the fetus original image that generates is carried out adopting the fixed threshold classification method that ultrasonoscopy is carried out image segmentation after the image filtering denoising pretreatment, find out edge wheel profile again by the sex of foetus position.
4. method as claimed in claim 2, it is characterized in that: among the step a, the fetus original image that generates is carried out after the image filtering denoising pretreatment, adopt the adaptive threshold selection method of supervised learning technology that ultrasonoscopy is carried out image segmentation, find out edge wheel profile again by the sex of foetus position.
5. the method for claim 1 is characterized in that: the extracting mode of the Primary Location point described in the step a is for extracting the curvature extreme point of object edge contour.
6. the method for claim 1 is characterized in that: the eigenvalue selection mode of the Primary Location point described in the step b is the continuous curvature value of choosing through the body outline of Primary Location point.
7. the method for claim 1 is characterized in that: the eigenvalue selection mode of the Primary Location point described in the step b is the spatial domain intensity profile in the topography zone at center for selecting with the Primary Location point, comprises average, variance information.
8. the method for claim 1 is characterized in that: the eigenvalue selection mode of the Primary Location point described in the step b is the picture element gray scale in the topography zone at center for selecting with the Primary Location point, and carries out the pyramid sample process.
9. the method for claim 1 is characterized in that: the grader that adopts in the described multiple Classifiers Combination structure is BP neural network classifier or weighting ballot grader.
10. the method for claim 1, it is characterized in that: the shielded image described in the step c is a mosaic.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CNB2007100759017A CN100462054C (en) | 2007-07-06 | 2007-07-06 | Method for shielding sex part on foetus image for preventing recognizing foetus sex |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CNB2007100759017A CN100462054C (en) | 2007-07-06 | 2007-07-06 | Method for shielding sex part on foetus image for preventing recognizing foetus sex |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN101081168A CN101081168A (en) | 2007-12-05 |
| CN100462054C true CN100462054C (en) | 2009-02-18 |
Family
ID=38911104
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CNB2007100759017A Active CN100462054C (en) | 2007-07-06 | 2007-07-06 | Method for shielding sex part on foetus image for preventing recognizing foetus sex |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN100462054C (en) |
Families Citing this family (18)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102048557A (en) * | 2009-10-26 | 2011-05-11 | 皇家飞利浦电子股份有限公司 | Device capable of preventing the abuse of ultrasonic image inspection and method thereof |
| CN101919709B (en) * | 2010-08-05 | 2012-08-22 | 珠海仁威医疗科技有限公司 | B ultrasonic all-in-one machine of pregnancy management and reproductive health services |
| CN104394771B (en) * | 2012-06-04 | 2017-07-04 | 泰尔哈绍梅尔医学研究基础设施和服务有限公司 | Ultrasonoscopy treatment |
| CN102783972B (en) * | 2012-08-21 | 2014-07-23 | 薛新华 | System for monitoring identification of gender of fetus without medical needs |
| JP5765583B2 (en) * | 2012-10-26 | 2015-08-19 | カシオ計算機株式会社 | Multi-class classifier, multi-class classifying method, and program |
| US20150342560A1 (en) * | 2013-01-25 | 2015-12-03 | Ultrasafe Ultrasound Llc | Novel Algorithms for Feature Detection and Hiding from Ultrasound Images |
| AU2013400936B2 (en) * | 2013-09-20 | 2017-08-03 | Transmural Biotech, S. L. | Image analysis techniques for diagnosing diseases |
| CN103610473B (en) * | 2013-11-21 | 2015-04-22 | 海信集团有限公司 | Ultrasonic image adaptive positioning measuring method and device |
| US10709416B2 (en) * | 2015-06-30 | 2020-07-14 | Wisconsin Alumni Research Foundation | Obstetrical imaging at the point of care for untrained or minimally trained operators |
| CN105894037B (en) * | 2016-04-21 | 2019-02-26 | 北京航空航天大学 | A kind of full supervised classification method of remote sensing images extracted based on SIFT training sample |
| CN106108950A (en) * | 2016-07-14 | 2016-11-16 | 宁波美童智能科技有限公司 | A kind of intelligent wireless ultrasonic fetal imaging screen method |
| CN107220652B (en) * | 2017-05-31 | 2020-05-01 | 北京京东尚科信息技术有限公司 | Method and device for processing pictures |
| CN107665333A (en) * | 2017-08-28 | 2018-02-06 | 平安科技（深圳）有限公司 | A kind of indecency image identification method, terminal, equipment and computer-readable recording medium based on convolutional neural networks |
| CN108111783B (en) * | 2017-12-22 | 2020-08-14 | 河南英富迪光电科技有限公司 | Outdoor self-charging image recognition infrared imager |
| CN109077720B (en) * | 2018-07-05 | 2021-03-23 | 广州视源电子科技股份有限公司 | Signal processing method, device, equipment and storage medium |
| CN109602452A (en) * | 2018-12-06 | 2019-04-12 | 余姚市华耀工具科技有限公司 | Organ orientation blocks system |
| CN109602453B (en) * | 2018-12-07 | 2021-11-16 | 徐梅 | Fetus pattern fuzzification processing platform |
| CN111820948B (en) * | 2019-04-20 | 2022-03-18 | 无锡祥生医疗科技股份有限公司 | Fetal growth parameter measuring method and system and ultrasonic equipment |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US4986274A (en) * | 1984-10-19 | 1991-01-22 | Stephens John D | Fetal anatomic sex assignment by ultrasonography during early pregnancy |
| CN2643833Y (en) * | 2003-08-27 | 2004-09-29 | 昆明阳光医疗科技有限公司 | Integrated complete-shielded ultrasound Doppler probe for fetal rhythm |
| CN1915177A (en) * | 2005-08-17 | 2007-02-21 | 北京天惠华数字技术有限公司 | Method for recording identification of fetus sexes through B ultrasonic |
-
2007
- 2007-07-06 CN CNB2007100759017A patent/CN100462054C/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US4986274A (en) * | 1984-10-19 | 1991-01-22 | Stephens John D | Fetal anatomic sex assignment by ultrasonography during early pregnancy |
| CN2643833Y (en) * | 2003-08-27 | 2004-09-29 | 昆明阳光医疗科技有限公司 | Integrated complete-shielded ultrasound Doppler probe for fetal rhythm |
| CN1915177A (en) * | 2005-08-17 | 2007-02-21 | 北京天惠华数字技术有限公司 | Method for recording identification of fetus sexes through B ultrasonic |
Also Published As
| Publication number | Publication date |
|---|---|
| CN101081168A (en) | 2007-12-05 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN100462054C (en) | Method for shielding sex part on foetus image for preventing recognizing foetus sex | |
| CN107833220B (en) | Fabric defect detection method based on deep convolutional neural network and visual saliency | |
| CN104809740B (en) | Knee cartilage image automatic segmentation method based on SVM and Hookean region growth | |
| JP6467041B2 (en) | Ultrasonic diagnostic apparatus and image processing method | |
| CN104715238B (en) | A kind of pedestrian detection method based on multi-feature fusion | |
| CN109345508A (en) | A kind of Assessing Standards For Skeletal method based on two stages neural network | |
| CN108022233A (en) | A kind of edge of work extracting method based on modified Canny operators | |
| CN105232081A (en) | Medical ultrasound assisted automatic diagnosis device and medical ultrasound assisted automatic diagnosis method | |
| CN105809175A (en) | Encephaledema segmentation method and system based on support vector machine algorithm | |
| CN109363698A (en) | A kind of method and device of breast image sign identification | |
| CN109363699A (en) | A kind of method and device of breast image lesion identification | |
| CN201076477Y (en) | Type-B ultrasonic device for fetus gender shielding | |
| CN108122221A (en) | The dividing method and device of diffusion-weighted imaging image midbrain ischemic area | |
| CN106156758B (en) | A kind of tidal saltmarsh method in SAR seashore image | |
| CN101615292A (en) | Human eye accurate positioning method based on half-tone information | |
| CN109003275A (en) | The dividing method of weld defect image | |
| CN112861723B (en) | Sports action recognition counting method and device based on human body gesture recognition and computer readable storage medium | |
| CN108875741A (en) | It is a kind of based on multiple dimensioned fuzzy acoustic picture texture characteristic extracting method | |
| CN108053401A (en) | A kind of B ultrasound image processing method and device | |
| CN100500100C (en) | Method and apparatus for recognizing intelligently ultrasound womb contraceptive ring image | |
| CN109102503A (en) | It is a kind of based on color space smoothly and improve the significant model of frequency tuning high score image change detection method | |
| CN112568933B (en) | Ultrasonic imaging method, apparatus and storage medium | |
| Samhitha et al. | Prediction of lung cancer using convolutional neural network (CNN) | |
| CN115546232A (en) | Liver ultrasonic image working area extraction method and system and electronic equipment | |
| TW201726064A (en) | Medical image processing apparatus and breast image processing method thereof |
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 |