CN102034232B - Method for segmenting medical image graph - Google Patents
Method for segmenting medical image graph Download PDFInfo
- Publication number
- CN102034232B CN102034232B CN 200910196422 CN200910196422A CN102034232B CN 102034232 B CN102034232 B CN 102034232B CN 200910196422 CN200910196422 CN 200910196422 CN 200910196422 A CN200910196422 A CN 200910196422A CN 102034232 B CN102034232 B CN 102034232B
- Authority
- CN
- China
- Prior art keywords
- image
- area
- region
- zone
- segmentation
- 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.)
- Expired - Fee Related
Links
Images
Landscapes
- Magnetic Resonance Imaging Apparatus (AREA)
- Apparatus For Radiation Diagnosis (AREA)
Abstract
The invention relates to a method for segmenting a medical image graph, which comprises the following steps: providing an image; adjusting the digital image into squares by using the pixel dimension as the unit; segmenting the image to determine the interested region; separating the square region of the interested region from the original image; and executing the steps of image segmentation and separation again for the separated square region to enhance the segmentation effect. The primary segmentation is carried out on the graph by using the method for segmenting a medical image graph, the position of the interested region is determined, and the repeated iteration segmentation is carried out on the position to obtain a better effect. The method does not depend on the specific human tissue and the specific image source, and can be used for segmentation for any grayscale image; and therefore, the invention has the advantage of universality.
Description
[technical field]
The present invention relates to the medical image processing field, relate in particular to a kind of dividing method of medical image figure.
[background technology]
In the medical image processing field, often need to adopt computer that image is separated, to obtain Useful Information and to carry out the auxiliary graphical analysis of computer.Existing image Segmentation Technology is divided into based on the dividing method in zone with based on the dividing method at edge.
The existing image processing method in this field comprises:
(1) Chinese patent " a kind of image partition method based on region growing and ant colony clustering " (application number 200710018667.4) has disclosed a kind of image partition method based on region growing and ant colony clustering, adopts the medium filtering in neighbours territory to remove filtering noise; The pixel that selection has the gray-scale value maximum carries out region growing as the seed point; Utilize spatial information and the half-tone information mentioned behind the region growing to define a kind of new guidance function then, use and between the zone, carry out the cluster merging in the ant group algorithm, obtain final segmentation result.
(2) Chinese patent " based on the medical image cutting method of horizontal collection and watershed method " (application number 02126555.0) has disclosed a kind of medical image cutting method based on level set and Watershed method, comprises the anisotropic diffusion filtering of removing noise; Adopt the Watershed method, image is carried out over-segmentation; The foundation of heap data structure is used for the net point that the arrowband, location has minimum time T; Adopt the FastMarching method, image is carried out last cutting apart.Be applied to CT or the MR image must be cut apart.
The shortcoming of prior art is that the problem that medical image segmentation exists is also not have a kind of general dividing method can solve all problems.Each method has its characteristics and the scope of application.
[summary of the invention]
Technical matters to be solved by this invention is, a kind of general dividing method is provided, can cut apart by the not simple image for a certain class image or some tissues, but a kind of image partition method with universality.
In order to address the above problem, the invention provides a kind of dividing method of medical image figure, comprise the steps: to provide piece image; Be unit with the Pixel Dimensions, it is 2 that digital picture is adjusted into the length of side
nIndividual pixel square, described n is natural number; Split image is to determine region-of-interest; The square area at this region-of-interest place is separated from original image; Carry out the above-mentioned step of cutting apart with separate picture again at the zone of separating, to improve segmentation effect.
As optional technical scheme, the type of described image is selected from a kind of in x-ray image, CT image, nuclear magnetic resonance image and the B ultrasonic image.
As optional technical scheme, the step of described split image further comprises: decompose image-region according to grey value difference, and will decompose the zone boundary of getting well and make marks; Deletion small size zone; Fill and the deletion tiny area; Calculate area and the ordering of each connected domain; Determine region-of-interest according to sorting position.
As optional technical scheme, described step according to grey value difference decomposition image-region further comprises: the allowed band of setting a gray consistency; Image is cut into four identical rectangular zones in length and breadth; Investigate the difference of the gray scale maximal value of divided area inside and minimum value whether in allowed band; If in allowed band, then stop cutting apart this zone; If not in allowed band, then continue to repeat above-mentioned segmentation step, till the difference of the gray scale maximal value of divided area inside and minimum value is in allowed band.
As optional technical scheme, the step in described deletion small size zone further comprises: set a minimum scale scope; The ON operation computing deletion area pixel area of application image and the ratio of entire image area are less than 8 connected domains of described minimum scale scope.
As optional technical scheme, the step of described filling and deletion tiny area further comprises: set one and fill radius value; Application image expands and erosion algorithm, and adopting a filling radius is the structure masterplate of described filling radius value, and edge hole hole and tiny area are filled and deleted.
As optional technical scheme, the described step of finding out the region-of-interest of determining sorting position further comprises: according to image type, determine the region area ordering in zone to be split; According to described region area ordering, in the sequencing information of each connected domain, find out region-of-interest.
The invention has the advantages that, adopt the dividing method of medical image figure tentatively to cut apart to figure, determine again to carry out repeatedly iterative cutting apart again at this position then and obtain better effect in the region-of-interest position.This method does not rely on concrete tissue and concrete image source, so long as gray level image can adopt this method to cut apart, therefore has universality.
[description of drawings]
It is the implementation step synoptic diagram of the specific embodiment of the present invention shown in the accompanying drawing 1
Accompanying drawing 2 to accompanying drawing 5 is to adopt the described method of the specific embodiment of the present invention to handle the decomposing schematic representation of a liver image;
It is the implementation result contrast synoptic diagram that adopts the described method of the specific embodiment of the present invention shown in the accompanying drawing 6.
[embodiment]
Elaborate below in conjunction with the embodiment of accompanying drawing to the dividing method of a kind of medical image figure provided by the invention.
Be the implementation step synoptic diagram of this embodiment shown in the accompanying drawing 1, comprise: step S100 provides piece image; Step S110 is unit with the Pixel Dimensions, and digital picture is adjusted into square; Step S121 decomposes image-region according to grey value difference, and will decompose the zone boundary of getting well and make marks; Step S122, deletion small size zone; Step S123 fills and the deletion tiny area; Step S124, area and the ordering of calculating each connected domain; Step S125 determines region-of-interest according to sorting position; Step S130 separates the square area at this region-of-interest place from original image; Step S140 at the square area of separating, carries out the above-mentioned step of cutting apart with separate picture, again to improve segmentation effect.
Refer step S100 provides piece image.The type of described image is selected from a kind of in the medium various common medical imagings of x-ray image, CT image, nuclear magnetic resonance image and B ultrasonic image.A liver image that provides in this embodiment is provided accompanying drawing 2.
Refer step S110 is unit with the Pixel Dimensions, and digital picture is adjusted into square.
Because the digital picture that obtains may be any irregular image, for the aspect on the subsequent treatment, this step will be unit with the Pixel Dimensions, and digital picture is adjusted into square, and its length of side is 2
nIndividual pixel (n is natural number) is so that follow-up image segmentation algorithm.To establish a capital be that the length of side is 2 because though present medical image differs
nThe rectangle of individual pixel, but Useful Information also all concentrates on the central area usually, so this step can not cause losing of useful information.
Reach shown in the accompanying drawing 3 and be the result after image shown in the accompanying drawing 2 is decomposed according to the method described above.
Refer step S121 decomposes image-region according to grey value difference, and will decompose the zone boundary of getting well and make marks.
A kind of optional implementation of this step is: set the allowed band of a gray consistency, for example the integral image gray scale is expressed 10% of scope; Image is cut into four identical rectangular zones in length and breadth; Investigate the difference of the gray scale maximal value of divided area inside and minimum value whether in allowed band; If in allowed band, then stop cutting apart this zone; If not in allowed band, then continue to repeat above-mentioned segmentation step, till the difference of the gray scale maximal value of divided area inside and minimum value is in allowed band.
Above step essence is to decompose image-region according to grey value difference, and will decompose the zone boundary of getting well and make marks, and make image seem at last to be made of a plurality of square grids, and the gray scale of each grid inside is similar.
Accompanying drawing 4 is depicted as and adopts above step that image shown in the accompanying drawing 3 is carried out result after the analyzing and processing.
Refer step S122, deletion small size zone.
A kind of optional implementation of this step is: set a minimum scale scope, this proportional range refers to the ratio of ON operation computing deletion area pixel area (pixel quantity) with the entire image area of application image, for example can be set to 5%; The ratio of the ON operation computing of application image deletion area pixel area and entire image area is less than 8 connected domains of described minimum scale scope (for example 5%).
Refer step S123 fills and the deletion tiny area.
A kind of optional implementation of this step is: set one and fill radius value, for example 8 pixels; Application image expands and erosion algorithm, and adopting a filling radius is the structure masterplate of described filling radius value, and edge hole hole and tiny area are filled and deleted.
Accompanying drawing 5 is depicted as and adopts above-mentioned steps that image shown in the accompanying drawing 4 is carried out result after the analyzing and processing.
Refer step S124, area and the ordering of calculating each connected domain.
The above-mentioned figure that disposes is divided into a plurality of UNICOMs zone according to gray consistency, calculates the area in each UNICOM zone respectively, and sorted according to area in each zone.
Refer step S125 determines region-of-interest according to sorting position.
A kind of optional implementation of this step is: according to image type, determine the region area ordering in zone to be split; According to the area ordering in described zone, in the sequencing information of each connected domain, find out region-of-interest.
Specifically, for the definite image (such as liver MR image) of a class, the ordering of the region area in zone to be split always determine (such as in liver MR image, the area sorting position of the hepatic region area of same sectional position in image determined), has ubiquity, therefore can determine the sorting position in zone to be split by image type, and region-of-interest is regarded as in the zone that will be in this sorting position.
Above step S121 is optional step to step S125, and its basic goal is split image, to determine region-of-interest.Shape, number, consistency criterion, deletion and the algorithm in filling small size zone and the algorithm of definite region-of-interest etc. that also can cut apart the back figure in other embodiment according to the characteristics adjustment of image are to obtain best segmentation effect.
Refer step S130 separates the square area at this region-of-interest place from original image.
The purpose of this step is to treat split image to carry out reorientation in original image.At first calculate the position coordinates P of the centre of form in original image tentatively cut apart the region-of-interest that obtains (x, y); Determine can comprise centered by this centre of form the size of the smallest square of region-of-interest again, the length of side is L; (x y) and the value of L, reorientates new square area from original image, this new square area should be the smallest square that can comprise region-of-interest in the original image according to P at last.
Refer step S140 at the square area of separating, carries out the above-mentioned step of cutting apart with separate picture, again to improve segmentation effect.
Repeated execution of steps S121 is to step S125, and last more strict parameter is compared in selection in the process of implementation, to obtain more accurate result.If the above-mentioned segmentation result afterwards of carrying out again still can not satisfy the demands, can also continue repeatedly to carry out above-mentioned steps, the judgement that stops of circulation is to cut apart figure and the expert is cut apart relatively deciding of figure according to current, and the expert is cut apart figure and referred to C1 in the example of similar back, C2, C3.That is to say, when the zone that obtains after cutting apart is the most similar to expert's cut zone, loop termination.For a certain class image to be split, also can obtain empirical parameter or cycle index by after an amount of study, the termination of circulation when controlling actual cutting apart.
Accompanying drawing 6 is depicted as the resulting result of liver segmentation who said method is applied to one group of nuclear magnetic resonance image, wherein three pictures of each row are identical, and A1, the A2 of first row and A3 figure adopt quaternary tree tentatively to cut apart the result who obtains, the B1 of secondary series, B2 and B3 figure are cut apart with the picture employing quaternary tree iteration of delegation, be the result that the described method of this embodiment obtains, tertial C1, C2 and C3 figure are that Senior Expert adopts the manual result of cutting apart in the industry, can be used as the segmentation result of standard, as a means of reference.The erroneous segmentation that liver upper right side (zone of irising out, the peninsula) exists among the figure A1 is corrected in A2; The zone of losing in B1 (upper right corner of irising out) obtains revising in B2; Over-segmentation zone in C1 (peninsula of irising out, right side) can be deleted in C2.
The above only is preferred implementation of the present invention; should be pointed out that for those skilled in the art, under the prerequisite that does not break away from the principle of the invention; can also make some improvements and modifications, these improvements and modifications also should be considered as protection scope of the present invention.
Claims (4)
1. the dividing method of a medical image figure is characterized in that, comprises the steps:
Piece image is provided;
Be unit with the Pixel Dimensions, it is 2 that digital picture is adjusted into the length of side
nIndividual pixel square, described n is natural number;
According to grey value difference split image zone, and will cut apart good zone boundary and make marks;
Set a minimum scale scope;
The ON operation computing deletion area pixel area of application image and the ratio of entire image area are less than 8 connected domains of described minimum scale scope;
Set one and fill radius value;
Application image expands and erosion algorithm, and adopting a filling radius is the structure masterplate of described filling radius value, and edge hole hole and tiny area are filled and deleted;
Calculate area and the ordering of each connected domain;
Determine region-of-interest according to sorting position;
The square area at this region-of-interest place is separated from original image;
Carry out the above-mentioned step of cutting apart with separate picture again at the square area of separating, to improve segmentation effect.
2. the dividing method of medical image figure according to claim 1 is characterized in that, the type of described image is selected from a kind of in x-ray image, CT image, nuclear magnetic resonance image and the B ultrasonic image.
3. the dividing method of medical image figure according to claim 1 is characterized in that, described step according to grey value difference split image zone further comprises:
Set the allowed band of a gray consistency;
Image is cut into four identical rectangular zones in length and breadth;
Investigate the difference of the gray scale maximal value of divided area inside and minimum value whether in allowed band;
If in allowed band, then stop cutting apart this zone;
If not in allowed band, then continue to repeat above-mentioned segmentation step, till the difference of the gray scale maximal value of divided area inside and minimum value is in allowed band.
4. the dividing method of medical image figure according to claim 1 is characterized in that, describedly determines that according to sorting position the step of region-of-interest further comprises:
According to image type, determine the region area ordering in zone to be split;
According to described region area ordering, in the sequencing information of each connected domain, find out region-of-interest.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN 200910196422 CN102034232B (en) | 2009-09-25 | 2009-09-25 | Method for segmenting medical image graph |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN 200910196422 CN102034232B (en) | 2009-09-25 | 2009-09-25 | Method for segmenting medical image graph |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN102034232A CN102034232A (en) | 2011-04-27 |
| CN102034232B true CN102034232B (en) | 2013-09-04 |
Family
ID=43887085
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN 200910196422 Expired - Fee Related CN102034232B (en) | 2009-09-25 | 2009-09-25 | Method for segmenting medical image graph |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN102034232B (en) |
Families Citing this family (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102646268B (en) * | 2012-02-22 | 2015-09-09 | 中国科学院自动化研究所 | The magnetic resonance image (MRI) brain structure automatic division method that Corpus--based Method multichannel chromatogram registration is optimized |
| CN102759617B (en) * | 2012-07-20 | 2014-06-25 | 无锡市东林科技发展有限责任公司 | Protein imprint automatic interpretation method and interpretation device thereof |
| CN104376556B (en) * | 2014-10-31 | 2017-06-16 | 四川大学 | A kind of rock CT images Target Segmentation method |
| WO2016086744A1 (en) | 2014-12-02 | 2016-06-09 | Shanghai United Imaging Healthcare Co., Ltd. | A method and system for image processing |
| CN105809656B (en) * | 2014-12-29 | 2020-03-31 | 上海联影医疗科技有限公司 | Medical image processing method and device |
| CN107507201A (en) * | 2017-09-22 | 2017-12-22 | 深圳天琴医疗科技有限公司 | A kind of medical image cutting method and device |
| CN107808384A (en) * | 2017-11-21 | 2018-03-16 | 江南大学 | Based on the local watershed fuzzy clustering method of gridding |
| CN109102514A (en) * | 2018-08-16 | 2018-12-28 | 广东工业大学 | A kind of image partition method, device, equipment and computer readable storage medium |
| CN110596117A (en) * | 2019-08-15 | 2019-12-20 | 山东科技大学 | Hyperspectral imaging-based rapid nondestructive detection method for apple surface damage |
| CN110610491B (en) * | 2019-09-17 | 2021-11-19 | 湖南科技大学 | Liver tumor region segmentation method of abdominal CT image |
| CN111195912B (en) * | 2020-01-08 | 2021-06-15 | 杭州未名信科科技有限公司 | Method and device for drawing portrait by using mechanical arm, robot and storage medium |
| CN111297344A (en) * | 2020-03-24 | 2020-06-19 | 陈恬慧 | Pulse-taking visual array acquisition method |
| CN112545566A (en) * | 2020-12-07 | 2021-03-26 | 武汉中旗生物医疗电子有限公司 | Fetal neck diaphragmatic layer measuring method and device |
| CN114037774B (en) * | 2022-01-10 | 2022-03-08 | 雅安市人民医院 | Method and device for sequencing and transmitting images of cross sections of cranium and brain and storage medium |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1471034A (en) * | 2002-07-24 | 2004-01-28 | 中国科学院自动化研究所 | Medical image segmentation method based on horizontal collection and watershed method |
| CN1489114A (en) * | 2003-09-09 | 2004-04-14 | 北京交通大学 | Diveided image coding and decoding method with arbitrary shape region segmentation |
| CN101286199A (en) * | 2007-09-14 | 2008-10-15 | 西北工业大学 | Method of image segmentation based on area upgrowth and ant colony clustering |
-
2009
- 2009-09-25 CN CN 200910196422 patent/CN102034232B/en not_active Expired - Fee Related
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1471034A (en) * | 2002-07-24 | 2004-01-28 | 中国科学院自动化研究所 | Medical image segmentation method based on horizontal collection and watershed method |
| CN1489114A (en) * | 2003-09-09 | 2004-04-14 | 北京交通大学 | Diveided image coding and decoding method with arbitrary shape region segmentation |
| CN101286199A (en) * | 2007-09-14 | 2008-10-15 | 西北工业大学 | Method of image segmentation based on area upgrowth and ant colony clustering |
Also Published As
| Publication number | Publication date |
|---|---|
| CN102034232A (en) | 2011-04-27 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN102034232B (en) | Method for segmenting medical image graph | |
| CN101657840B (en) | System and method for cell analysis in microscopy | |
| CN104809723B (en) | The three-dimensional CT image for liver automatic division method of algorithm is cut based on super voxel and figure | |
| Carr et al. | Flexible isosurfaces: Simplifying and displaying scalar topology using the contour tree | |
| US6741755B1 (en) | System and method providing mixture-based determination of opacity | |
| CN109102506B (en) | Automatic segmentation method for abdominal CT liver lesion image based on three-level cascade network | |
| Hashemi et al. | A tomographic imagery segmentation methodology for three-phase geomaterials based on simultaneous region growing | |
| JP2007111531A (en) | Method for improving image consistency, signal, computer readable medium and system | |
| US20090252429A1 (en) | System and method for displaying results of an image processing system that has multiple results to allow selection for subsequent image processing | |
| US6704439B1 (en) | Method for identification of alveolar nerve region in mandible image | |
| CN101576997A (en) | Abdominal organ segmentation method based on secondary three-dimensional region growth | |
| US7394933B2 (en) | Region competition via local watershed operators | |
| CN103918004B (en) | The self adaptation application of metal artifacts reduction algorithm | |
| Kohler et al. | Automated compartmental analysis for high-throughput skeletal phenotyping in femora of genetic mouse models | |
| JP2005296605A (en) | Method of segmenting a radiographic image into diagnostically relevant and diagnostically irrelevant regions | |
| CN105719243A (en) | Image processing device and method | |
| CN103700085A (en) | Cutting method of pectoral muscle region in mammary gland X-ray image | |
| KR101423835B1 (en) | Liver segmentation method in a medical image | |
| Maitra et al. | Accurate breast contour detection algorithms in digital mammogram | |
| CN109345536B (en) | Image super-pixel segmentation method and device | |
| CN103544695A (en) | Efficient medical image segmentation method based on game framework | |
| CN109685783A (en) | A kind of method for cell count based on skeletal extraction | |
| CN111292310B (en) | Method, device, electronic equipment and storage medium for acquiring biological tissue image | |
| CN108596840A (en) | A kind of data set Enhancement Method for deep learning evaluation blood vessel network developmental level | |
| CN112348826A (en) | Interactive liver segmentation method based on geodesic distance and V-net |
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 | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20130904 Termination date: 20140925 |
|
| EXPY | Termination of patent right or utility model |