CN101085364A - Method for detecting mammary cancer armpit lymph gland transferring focus - Google Patents
Method for detecting mammary cancer armpit lymph gland transferring focus Download PDFInfo
- Publication number
- CN101085364A CN101085364A CN 200610046830 CN200610046830A CN101085364A CN 101085364 A CN101085364 A CN 101085364A CN 200610046830 CN200610046830 CN 200610046830 CN 200610046830 A CN200610046830 A CN 200610046830A CN 101085364 A CN101085364 A CN 101085364A
- Authority
- CN
- China
- Prior art keywords
- vote
- interest
- edge
- mammary cancer
- breast
- 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
Landscapes
- Image Analysis (AREA)
- Apparatus For Radiation Diagnosis (AREA)
Abstract
The invention disclose a detection method of axilla lymph node metastasis which comprises the following steps: extracting zone of interest from X-ray diagram of inner side mammary gland in oblique position, performing edge extract of each zone of interest to obtain two-value outline map, detecting multiple round or similar round edge with different semidiameter through Hough transformation, selecting to remove parametric combination not correspondent with round or similar round edge and reserve round edge determinant region corresponding to various parametric combination which is the lymph node. The inventive method has good detecting accuracy, broad application scope, and good controllability.
Description
Technical field
The present invention relates to a kind of medical science detection technique, specifically a kind of method for detecting mammary cancer armpit lymph gland transferring focus.
Background technology
Breast carcinoma is malignant tumor common, high incidence, serious threat women's health.When the lump diameter during greater than 2cm breast cancer cell just might shift.Breast cancer cell is often transferred to the axillary gland of homonymy earlier, and the lymph node quantity that transfer takes place is many more, and patient's survival rate is low more.
In the prior art, the breast X-ray image be conventional also be the most effective breast cancer diagnosis foundation.In clinical practice to the breast carcinoma diagnosis and treatment, the universal method of obtaining the breast X-ray image is that the bilateral breast is carried out position (Cranio Caudal end to end, CC) and interior lateral oblique position (Mediolateral Oblique, MLO) irradiation, obtain the bilateral breast radioscopic image.Armpit lymph gland transferring focus only is included in the breast X-ray image of MLO position.The lymph no metastatic lesion shows as one or more high density shades on the breast X-ray image, rounded or sub-circular, and along with the growth of lymph node, a plurality of lymph nodes may stick together.
In the prior art, based on the breast X-ray image, utilize computer technology still blank to the method that mammary cancer armpit lymph gland transferring focus detects.
The Hough conversion is a kind of image processing method commonly used, is used for the edge of the given shape that detected image comprises, as straight line, circular edge etc.The Hough conversion is based on the voting principle, and its process is: determine to describe the parameter of edge shape, transformation space is formed in the combination of these parameters; Every bit on the edge is voted transformation space, wins the parameter combinations of majority vote in theory and wins.
In the practical application of Flame Image Process, the problem of Hough conversion is how accurately to determine promptly to need to determine the triumph condition to such an extent that how many tickets be can be regarded as triumph; In addition, in actual detection is used, simple application Hough conversion often can not reach ideal detection effect, auxiliary other detection meanss that provides are provided, such as when detecting a plurality of edge simultaneously (as detecting the different a plurality of circles of length different many straight lines, radius etc. simultaneously), number of votes obtained can not be as unique triumph foundation, therefore can not detect circle or sub-circular edge in the image very soon with the Hough conversion of routine.
Summary of the invention
In order to remedy above-mentioned deficiency, the object of the present invention is to provide a kind of the detection accurately and method for detecting mammary cancer armpit lymph gland transferring focus with controllability.
For achieving the above object, the technical solution used in the present invention is:
A. the breast X-ray image to interior lateral oblique position position carries out the area-of-interest extraction; B. above extracted each area-of-interest is carried out edge extracting, obtain binary edge map; C. on the binary edge map of each area-of-interest, pass through the different circular or similar circular edge of a plurality of radiuses of Hough change detection; D. select in the result of the Hough conversion that above-mentioned steps c obtains, do not remove and make up with circular or similar circular edge corresponding parameter, the determined zone of circular edge of each parameter combinations correspondence that remains is detected lymph node.
The breast X-ray image of described internal lateral oblique position position carries out area-of-interest and extracts, and the steps include: that a1. carries out arm-pit areas and extracts on the breast X-ray image of interior lateral oblique position position; A2. above extracted arm-pit areas being carried out area-of-interest extracts;
The described Hough of step c is transformed to and passes through formula: Vote
Temp=Vote
Org/ 2 π R, Vote
Unitary=Vote
Temp/ max (Vote
Temp) the original number of votes obtained of all parameter combinations is handled Vote wherein
OrgBe original number of votes obtained; Vote
TempIntermediate value for processing procedure; Vote
UnitaryBe the number of votes obtained after handling; Described screening technique is: to each parameter combinations, choose inside and outside two zones of the circle of its correspondence, ask the average gray in these two zones respectively, if the average gray difference of the two is poor greater than predefined reference gray level, corresponding circular edge of this parameter combinations then; Otherwise not corresponding, remove; The present invention also has step e: labelling testing result on image, soon external pictorial display is drawn on image in detected lymph node zone in the steps d, wherein, the external figure in lymph node zone then is merged into a figure with these external figures and shows on image if two or more overlapping arranged.
The present invention has the following advantages and beneficial effect:
1. focus detects accurately.The present invention is applied on the breast X-ray image of clinical collection,, can detects axillary gland accurately, filled up the blank of method for detecting mammary cancer armpit lymph gland transferring focus owing to adopted improved Hough alternative approach;
2. applied widely.Utilize original image information to detecting the method that PRELIMINARY RESULTS is further checked because the present invention has adopted, those ND lymph nodes, the lymph node that sticks together are also had good detection effect, enlarged the scope of application of Clinical detection focus; And owing to can remove the hollow lymph node that does not characterize breast carcinoma, thereby further improved the detection accuracy rate;
3. has good controllability.Radius R is changeable parameter in the parameter combinations of Hough conversion of the present invention, because each parameter combinations is corresponding with circle or sub-circular edge that a last lymph node produces, therefore by changing the scope of radius R, can regulate the radius that detects lymph; The reference gray level difference also is the parameter that can change, and can regulate according to the use preference of user.
Description of drawings
Fig. 1 is lateral oblique position breast X-ray image in the present invention;
Fig. 2 (a) is for carrying out the image after arm-pit areas is extracted;
Fig. 2 (b) is the binary edge map of each ROI in the arm-pit areas;
Interior exterior domain in Fig. 3 parameter combinations selection method is chosen sketch map;
Fig. 4 (a)～(d) is a design sketch of the present invention;
Fig. 5 is the inventive method flow chart;
Fig. 6 is the workflow diagram of one embodiment of the invention.
The specific embodiment
As Fig. 5, shown in Figure 6, the following method that provides is provided the armpit lymph gland transferring focus of the breast X-ray image of lateral oblique position in the present invention is directed to:
A. by the patient being taken the breast X-ray image of breast X-ray Image Acquisition interior lateral oblique position as shown in Figure 1.
B. carrying out arm-pit areas on the breast X-ray image of interior lateral oblique position extracts; Present embodiment employing iteration threshold method is asked the optimal segmenting threshold between mammary gland zone and background, and utilizes this optimal segmenting threshold to extract arm-pit areas.
C. above extracted arm-pit areas is carried out region of interest and extract, obtain a plurality of region of interest, specifically comprise: ask the grey level histogram of arm-pit areas, getting the brightest one part of pixel corresponding gray scale value of this rectangular histogram is reference threshold; With the point of all gray values in the arm-pit areas greater than reference threshold, according to its position and with the distance of other point, be divided into some groups, get the boundary rectangle of the point of each group, be region of interest (Region Of Interest, ROI).
D. above extracted each region of interest is carried out edge extracting, and the passing threshold conversion obtains binary edge map, specifically comprise: each ROI that is extracted among the step c is carried out rim detection respectively, present embodiment adopts the sobel operator, can also adopt robert operator, prewitt operator, Gauss's Laplace operator etc.; The computational transformation threshold value, promptly to each ROI, the result that above-mentioned rim detection is obtained asks grey level histogram, and it is the conversion threshold value that present embodiment is got the darkest one part of pixel corresponding gray of this grey level histogram; Threshold transformation promptly carries out threshold transformation according to above-mentioned conversion threshold value, obtains the binary edge map of this ROI;
Shown in Fig. 2 (a), Fig. 2 (b), the binary edge map that is respectively the arm-pit areas image of extraction and arm-pit areas is carried out obtain behind ROI extraction, rim detection, the threshold transformation.
E. to the binary edge map of each region of interest with improved Hough change detection circular or similar circular edge wherein, improved Hough conversion, promptly in linear transformation, can detect the Hough conversion at different circle of a plurality of radiuses or sub-circular edge simultaneously, specifically comprise: determine parameter, with Hough change detection circle or sub-circular edge, the parameter that present embodiment is determined comprises the position (x in the center of circle
c, y
c) and radius of circle R, transformation space is formed in the combination of these parameters, the corresponding parameter combinations of the every bit in the transformation space; To the every bit in the transformation space, promptly parameter combinations is voted; Because aforesaid Hough conversion process is not considered the influence that the radius difference is brought, in transformation space, it is many that often real than the minor radius zone circular edge corresponding parameters of the number of votes obtained of non-circular edge, long radius zone corresponding parameters combination makes up number of votes obtained, make the combination of real circular edge corresponding parameters can not be in ballot " triumphs " or " triumph " advantage not obvious, improvement Hough conversion: the number of votes obtained of all parameter combinations is handled according to formula (1), (2):
Vote
temp＝Vote
org/2πR (1)
Vote
unitaty＝Vote
temp/max(Vote
temp) (2)
Wherein, Vote
OrgBe original number of votes obtained; Vote
UnitaryBe the number of votes obtained after handling; Vote
TempIntermediate value for processing procedure; The effect of formula (1) is to consider the different influence of radius; The effect of formula (2) is that number of votes obtained is carried out normalized, and the final number of votes obtained that makes each parameter combinations is in [0,1] scope; F. select in the result of the Hough conversion that above-mentioned steps e obtains, do not remove and make up with circular or similar circular edge corresponding parameter, the determined zone of circular edge of each parameter combinations correspondence that remains is detected lymph node; The problem of Hough conversion is to determine that how many tickets are parameter combinations get and just can be regarded as triumph, for avoiding omission, parameter combinations as much as possible should be can be regarded as " triumph ", utilize original image information to judge the whether corresponding circular edge of this parameter combinations then, as not corresponding, then remove this parameter combinations; Judge parameter combinations that the Hough conversion obtains whether with the corresponding method of a circle or sub-circular two-value edge as shown in Figure 3, choose the inside of the circle of parameter combinations correspondence, outside two zones, wherein 1 is the corresponding circular edge interior zone of parameter combinations, 2 is the corresponding circular edge of parameter combinations perimeter, and the center of circle of circular edge inside, perimeter 1,2 is the (x of parameter combinations in the present embodiment
c, y
c) component, circular edge interior zone 1 is got the circle that radius is R/2, the annulus that radius is 1.2R～1.4R is got in circular edge perimeter 2, ask the average gray in these two zones respectively, if the average gray difference of the two is poor greater than predefined reference gray level, corresponding circular edge of this parameter combinations then, otherwise be not; The performance of lymph node on the breast X-ray image has hollow and solid two kinds, and solid is cancer metastasis, and hollow is not, should remove, and uses this determination methods, can also remove the circular edge of hollow lymph node correspondence;
H. show testing result: through above steps, the corresponding detected lymph node of each parameter combinations that stays.The external figure of each parameter combinations correspondence is signed on the image, marked detected lymph node.If there are two or more external figures to overlap, may be due to two or more lymph gland transferring focus stick together, then these figures are merged into a pictorial display; Shown in Fig. 4 (a)～(b), wherein Fig. 4 (a) shows the result that typical armpit lymph gland transferring focus is detected, the result that Fig. 4 (b) detects the armpit lymph gland transferring focus that is difficult for explanation, Fig. 4 (c) shows the result that the armpit lymph gland transferring focus that sticks together is detected, Fig. 4 (d) demonstration detects solid lymph node, and removed the hollow lymph node shown in the white edge, above-mentioned each display result is all indicated with square frame.
The inventive method can not only provide foundation for the doctor determines the quantity and the position of the axillary gland that the generation focus shifts, and owing to benign tumor does not shift, so help to judge the very pernicious of tumor.
Claims (5)
1. method for detecting mammary cancer armpit lymph gland transferring focus is characterized in that:
A. the breast X-ray image to interior lateral oblique position position carries out the area-of-interest extraction;
B. above extracted each area-of-interest is carried out edge extracting, obtain binary edge map;
C. on the binary edge map of each area-of-interest, pass through the different circular or similar circular edge of a plurality of radiuses of Hough change detection;
D. select in the result of the Hough conversion that above-mentioned steps c obtains, do not remove and make up with circular or similar circular edge corresponding parameter, the determined zone of circular edge of each parameter combinations correspondence that remains is detected lymph node.
2. by the described method for detecting mammary cancer armpit lymph gland transferring focus of claim 1, it is characterized in that: the breast X-ray image of described internal lateral oblique position position carries out area-of-interest and extracts, and the steps include:
A1. carrying out arm-pit areas on the breast X-ray image of interior lateral oblique position position extracts;
A2. above extracted arm-pit areas being carried out area-of-interest extracts;
3. by the described method for detecting mammary cancer armpit lymph gland transferring focus of claim 1, it is characterized in that the described Hough of step c is transformed to passes through formula: Vote
Temp=Vote
Org/ 2 π R, Vote
Unitary=Vote
Temp/ max (Vote
Temp) the original number of votes obtained of all parameter combinations is handled Vote wherein
OrgBe original number of votes obtained; Vote
TempIntermediate value for processing procedure; Vote
UnitaryBe the number of votes obtained after handling.
4. by the described method for detecting mammary cancer armpit lymph gland transferring focus of claim 1, it is characterized in that: described screening technique is: to each parameter combinations, choose inside and outside two zones of the circle of its correspondence, ask the average gray in these two zones respectively, if the average gray difference of the two is poor greater than predefined reference gray level, corresponding circular edge of this parameter combinations then; Otherwise not corresponding, remove.
5. by the described method for detecting mammary cancer armpit lymph gland transferring focus of claim 1, it is characterized in that also having step e: labelling testing result on image, soon external pictorial display is drawn on image in detected lymph node zone in the steps d, wherein, the external figure in lymph node zone then is merged into a figure with these external figures and shows on image if two or more overlapping arranged.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN 200610046830 CN101085364A (en) | 2006-06-07 | 2006-06-07 | Method for detecting mammary cancer armpit lymph gland transferring focus |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN 200610046830 CN101085364A (en) | 2006-06-07 | 2006-06-07 | Method for detecting mammary cancer armpit lymph gland transferring focus |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN101085364A true CN101085364A (en) | 2007-12-12 |
Family
ID=38936646
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN 200610046830 Pending CN101085364A (en) | 2006-06-07 | 2006-06-07 | Method for detecting mammary cancer armpit lymph gland transferring focus |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN101085364A (en) |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101609558B (en) * | 2009-07-15 | 2011-09-07 | 宁波大学 | Preprocessing method for breast region extraction in breast molybdenum target X-ray image |
| CN103425986A (en) * | 2013-08-31 | 2013-12-04 | 西安电子科技大学 | Breast lump image feature extraction method based on edge neighborhood weighing |
| CN104657984A (en) * | 2015-01-28 | 2015-05-27 | 复旦大学 | Automatic extraction method of three-dimensional breast full-volume image regions of interest |
| CN106127216A (en) * | 2016-03-16 | 2016-11-16 | 青岛大学 | A kind of method for detecting mammary cancer armpit lymph gland transferring focus |
| CN106780436A (en) * | 2016-11-18 | 2017-05-31 | 北京郁金香伙伴科技有限公司 | A kind of medical imaging display parameters determine method and device |
| CN108921821A (en) * | 2018-06-01 | 2018-11-30 | 中国人民解放军战略支援部队信息工程大学 | Method of discrimination based on the LASSO mammary cancer armpit lymph gland transfering state returned |
| CN109846513A (en) * | 2018-12-18 | 2019-06-07 | 深圳迈瑞生物医疗电子股份有限公司 | Ultrasonic imaging method, system and image measuring method, processing system and medium |
-
2006
- 2006-06-07 CN CN 200610046830 patent/CN101085364A/en active Pending
Cited By (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101609558B (en) * | 2009-07-15 | 2011-09-07 | 宁波大学 | Preprocessing method for breast region extraction in breast molybdenum target X-ray image |
| CN103425986A (en) * | 2013-08-31 | 2013-12-04 | 西安电子科技大学 | Breast lump image feature extraction method based on edge neighborhood weighing |
| CN103425986B (en) * | 2013-08-31 | 2016-08-10 | 西安电子科技大学 | Mammary gland tumor image characteristic extracting method based on edge neighborhood weighting |
| CN104657984A (en) * | 2015-01-28 | 2015-05-27 | 复旦大学 | Automatic extraction method of three-dimensional breast full-volume image regions of interest |
| CN104657984B (en) * | 2015-01-28 | 2018-10-16 | 复旦大学 | The extraction method of three-D ultrasonic mammary gland total volume interesting image regions |
| CN106127216A (en) * | 2016-03-16 | 2016-11-16 | 青岛大学 | A kind of method for detecting mammary cancer armpit lymph gland transferring focus |
| CN106780436A (en) * | 2016-11-18 | 2017-05-31 | 北京郁金香伙伴科技有限公司 | A kind of medical imaging display parameters determine method and device |
| CN106780436B (en) * | 2016-11-18 | 2020-08-25 | 北京郁金香伙伴科技有限公司 | Medical image display parameter determination method and device |
| CN108921821A (en) * | 2018-06-01 | 2018-11-30 | 中国人民解放军战略支援部队信息工程大学 | Method of discrimination based on the LASSO mammary cancer armpit lymph gland transfering state returned |
| CN109846513A (en) * | 2018-12-18 | 2019-06-07 | 深圳迈瑞生物医疗电子股份有限公司 | Ultrasonic imaging method, system and image measuring method, processing system and medium |
| CN109846513B (en) * | 2018-12-18 | 2022-11-25 | 深圳迈瑞生物医疗电子股份有限公司 | Ultrasonic imaging method, ultrasonic imaging system, image measuring method, image processing system, and medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN101085364A (en) | Method for detecting mammary cancer armpit lymph gland transferring focus | |
| CN102525381B (en) | The recording equipment of image processing apparatus, image processing method and embodied on computer readable | |
| CN105447872A (en) | Method for automatically identifying liver tumor type in ultrasonic image | |
| CN106991694B (en) | Based on marking area area matched heart CT and ultrasound image registration method | |
| CN106780448A (en) | A kind of pernicious sorting technique of ultrasonic Benign Thyroid Nodules based on transfer learning Yu Fusion Features | |
| CN105261015B (en) | Eye fundus image blood vessel automatic division method based on Gabor filter | |
| CN105654490A (en) | Lesion region extraction method and device based on ultrasonic elastic image | |
| CN101238987A (en) | Processing method of CT cerebral hemorrhage image | |
| CN104751437B (en) | Lung's extraction method based on chest CT image | |
| CN105825516A (en) | DICOM image blood flow analysis system | |
| WO2002094098A1 (en) | Diagnostic feature extraction in dermatological examination | |
| Patil et al. | A review paper on brain tumor segmentation and detection | |
| CN107798679A (en) | Breast molybdenum target image breast area is split and tufa formation method | |
| CN102737379A (en) | Captive test (CT) image partitioning method based on adaptive learning | |
| CN108596894A (en) | A kind of prostate automatic Mesh Partition Method for multi-parameter nuclear magnetic resonance image | |
| Su et al. | Delineation of carpal bones from hand X-ray images through prior model, and integration of region-based and boundary-based segmentations | |
| CN102073872B (en) | Image-based method for identifying shape of parasite egg | |
| CN108268897A (en) | A kind of multi-modal multi-level computer aided diagnosing method of ultrasound of tumor of breast | |
| CN105427325A (en) | Automatic lung tumour segmentation method based on random forest and monotonically decreasing function | |
| WO2024067527A1 (en) | Hip joint angle measurement system and method | |
| CN103034979A (en) | Ultrasonic image definition improving method | |
| CN104881858A (en) | Method and device for extracting background parenchymal enhancement in breast | |
| CN106169084A (en) | A kind of SVM mammary gland sorting technique based on Gauss kernel parameter selection | |
| CN101425140A (en) | Encephalic angioma image recognizing and detecting method based on framework characteristic | |
| CN106780718A (en) | A kind of three-dimensional rebuilding method of paleontological fossil |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C02 | Deemed withdrawal of patent application after publication (patent law 2001) | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Open date: 20071212 |