CN114841958A - Method for automatically extracting image of brain subcutaneous tissue patch - Google Patents
Method for automatically extracting image of brain subcutaneous tissue patch Download PDFInfo
- Publication number
- CN114841958A CN114841958A CN202210477260.2A CN202210477260A CN114841958A CN 114841958 A CN114841958 A CN 114841958A CN 202210477260 A CN202210477260 A CN 202210477260A CN 114841958 A CN114841958 A CN 114841958A
- Authority
- CN
- China
- Prior art keywords
- brain
- image
- gray
- subcutaneous tissue
- tissue
- 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
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0012—Biomedical image inspection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/11—Region-based segmentation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/136—Segmentation; Edge detection involving thresholding
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/30—Determination of transform parameters for the alignment of images, i.e. image registration
- G06T7/33—Determination of transform parameters for the alignment of images, i.e. image registration using feature-based methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10028—Range image; Depth image; 3D point clouds
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30004—Biomedical image processing
- G06T2207/30016—Brain
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30004—Biomedical image processing
- G06T2207/30024—Cell structures in vitro; Tissue sections in vitro
Landscapes
- Engineering & Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Health & Medical Sciences (AREA)
- Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
- Radiology & Medical Imaging (AREA)
- Quality & Reliability (AREA)
- Magnetic Resonance Imaging Apparatus (AREA)
- Image Analysis (AREA)
Abstract
An automatic extraction method of a brain subcutaneous tissue patch image relates to the technical field of medical image processing. In brain image processing tasks, the size of the target tissue is usually much smaller than the size of the entire brain tissue, and especially in image registration, only pixels in the vicinity of the target tissue are valid, while the remaining pixels are invalid, and memory space and computation time are wasted. Aiming at the problem, the method provides an automatic extraction method of the patch image of the subcutaneous brain tissue, and the method can automatically extract the patch image containing all the subcutaneous brain tissue according to the position information and the gray distribution information of the subcutaneous brain tissue. By extracting the patch image from the whole image and processing only the patch image in the subsequent image processing operation, a large number of invalid pixels can be removed, the data amount required to be processed is reduced, and the computational complexity is reduced.
Description
Technical Field
The invention relates to the technical field of image processing, in particular to an automatic extraction method of a brain subcutaneous tissue patch image.
Background
In brain MR/CT image registration or segmentation tasks, the size of the target tissue is typically much smaller than the size of the entire brain tissue. Taking a transverse position view as an example, the area of the subcutaneous brain tissue in one image only occupies 1/30-1/100 of the area of the brain tissue in the image.
Image registration is one of the commonly used medical image processing techniques, and brain image registration can provide doctors with clearer brain tissue information, and is often used as a preprocessing step for brain tissue segmentation.
The method maps label information to a target image space through image registration, and under the condition of not considering image rotation and translation factors, the mapping range of pixel points of most nonlinear registration algorithms is limited and can only be mapped to an adjacent region. Therefore, in the registration of the two images, only the pixels in the adjacent area of the target tissue are valid, and the rest pixel points are not mapped to the target tissue space, so that the registration of the target tissue is invalid, and the calculation time is wasted.
Disclosure of Invention
The invention aims to provide an automatic extraction method of a brain subcutaneous tissue patch image, which extracts the patch image containing all brain subcutaneous tissues according to the position information and the gray distribution information of the brain subcutaneous tissues and aims to remove invalid image data, reduce the registration calculation time of the brain image and accelerate the registration speed.
The technical scheme adopted by the invention for solving the problems is as follows:
an automatic extraction method of a brain subcutaneous tissue patch image is characterized by comprising the following steps:
step one, setting the size of an initial search area and a central area of a brain subcutaneous tissue: the initial search area of the brain subcutaneous tissue and the size of the central area of the brain subcutaneous tissue are determined by analyzing the brain image dataset, and the specific process is as follows:
1) carrying out pixel value normalization processing on a gray level image in the brain image data set;
2) expanding all images in the brain image dataset to the same size by adopting a background pixel supplementing mode, and placing brain tissues in the images in the center of the images;
3) analyzing the boundary position of the brain subcutaneous tissue in the brain image data set, determining the left limit position, the right limit position, the upper limit position and the lower limit position, and setting a rectangular area as an initial search area of the brain subcutaneous tissue according to the four limit positions;
4) the gray distribution diagram of the brain subcutaneous tissue is approximately in normal distribution, the peak value in the gray distribution diagram is the characteristic gray value of the brain subcutaneous tissue, and the minimum rectangle size containing all the characteristic gray values is selected as the size of the central area of the brain subcutaneous tissue.
Step two, extracting an image gray distribution map: respectively extracting pixel gray distribution maps of a target image and a central region image of a brain subcutaneous tissue, wherein the specific method for extracting the image gray distribution maps comprises the following steps: carrying out pixel normalization processing on the image, dividing the pixel value of 0-1 into 20 intervals equally, counting the number of pixel points in each interval, and drawing a gray distribution graph by taking the pixel value as an abscissa and the number of the pixel points as an ordinate.
Step three, calculating the membership degree of the image of the brain subcutaneous tissue patch, wherein the specific process comprises the following steps:
1) calculating pixel point proportion information D of peak value and neighborhood in brain subcutaneous tissue gray level distribution diagram s ：
Where i is the interval number corresponding to the peak in the distribution diagram, and N i Is the number of pixel points in i in the interval, n is the total number of pixel points in the central region, beta 1 ，β 2 ，β 3 Is a weight coefficient, wherein 1 Representing the relative position of the target peak to the peak of the entire image, beta 2 As peak ratio information weight, beta 2 Pixel site proportion weight of peak value and its neighborhood, N j The number of pixel points in the interval j;
2) calculating the similarity M of each interval gradient information in the gray distribution map of the target image and each interval gradient information in the gray distribution map of the central region of the subcutaneous brain tissue s ：
In the formula S thr As gradient threshold, Ng j Gradient values of adjacent intervals in the gray distribution;
S thr obtained by analysis of the brain image dataset, in the following way:
a. extracting the gray distribution map of the gray image in the brain image data set and the gray distribution map of the central region of the brain subcutaneous tissue by adopting the method of the second step, and performing cubic B-spline optimization on the extracted gray distribution maps to avoid the interference of abnormal values;
b. calculating gradient value Ng of each interval j ；
Ng j ＝|N j+1 -N j |，j∈[0,19] (4)
NG＝[Ng 1 ,Ng 2 ,…,Ng 19 ] (5)
c. Fitting the gray distribution diagram of the gray image in the brain data set and the gradient variance of the gray distribution diagram of the target image, wherein the calculation formula is as follows:
in the formula, ANg j For gradient values of adjacent intervals in the gray profile of a gray image in a brain dataset, TNg j For adjacent in the gray distribution of the target imageGradient value of interval, a j For the weight coefficient, tn is the number of maps, and S is calculated by the formulas (6), (7) and (8) thr ；
3) Calculating the membership Ey of the target tissue sub-image:
Ey＝α 1 M s +α 2 D s (9)
in the formula of alpha 1 And alpha 2 Are weight coefficients.
Step four, extracting a brain subcutaneous tissue patch image, wherein the specific process is as follows: traversing the search space, searching for a central region of the brain subcutaneous tissue with the maximum membership value, and extracting a brain subcutaneous tissue patch image according to the central region of the brain subcutaneous tissue.
The invention has the beneficial effects that:
the invention provides an automatic extraction method of a brain subcutaneous tissue patch image, which can automatically extract the patch image containing all brain subcutaneous tissues according to the position information and the gray distribution information of the brain subcutaneous tissues.
The invention can remove invalid pixels outside the neighborhood of the brain subcutaneous tissue, reduce the time for registering and calculating the brain image and accelerate the registering speed.
Drawings
Fig. 1 is a flowchart of an automatic extraction method of a patch image of a brain subcutaneous tissue according to the present invention.
Fig. 2 is a schematic diagram of initial search position and sub-image center size setting.
FIG. 3 is a schematic diagram of the result of a human brain subcutaneous tissue patch image extracted by the method of the present invention.
Detailed Description
The present invention will be described in further detail with reference to the accompanying drawings and examples. It is to be understood that the specific embodiments described herein are merely illustrative of the invention and are not limiting of the invention. For the convenience of description, only some structures related to the present invention are shown in the drawings, not all structures, and for the convenience of illustration, the schematic diagrams are only examples, and should not limit the scope of the present invention. As shown in fig. 1, fig. 1 is a flowchart of an embodiment of a method for automatically extracting a patch image of a subcutaneous brain tissue according to the present invention, and the method includes:
data selection: selecting IXI a brain image dataset (http:// brain-level. org/brain-atlas /), the dataset comprising 30 sets of three-dimensional human brain MR atlases, wherein each set of atlases comprises an atlas gray-scale image and an atlas label image of the target tissue marked by the physician; one set of maps was selected as the target map and the others as the test maps.
In this example, 6 subcutaneous tissues of the brain were extracted, respectively: thalamus, hippocampus, caudate nucleus, putamen, globus pallidus, and amygdala.
The method comprises the following specific steps of automatically extracting the brain subcutaneous tissue patch image:
step one, setting an initial search position of a brain subcutaneous tissue and the size of a central area of a patch image:
a transection image was selected for testing and the initial search area and central area sizes and expanded sizes were set by analysis of the subcutaneous brain tissue in the IXI brain image dataset, as shown in fig. 2.
The parameters obtained after analysis are shown in table 1, in which the parameters Lw and Rw represent the left and right limit positions of the initial search area; lh and Hh represent the upper and lower limit positions of the initial search region, and Fd is the distance between the peak of the target tissue gray profile and the peak of the whole image gray profile.
TABLE 1 IXI data set Primary search area for subcutaneous brain tissue and Central area size parameters
Step two, extracting pixel gray distribution maps of the target image and the central region image of the brain subcutaneous tissue: equally dividing the intensity value of 0-1 into 20 intervals, extracting the gray distribution diagram of the target image and the central area image, and then carrying out cubic B-spline optimization on the gray distribution diagram.
Step three, calculating the membership degree of the image of the brain subcutaneous tissue patch
1) S of each image is calculated by equation (6) T Value and calculating the threshold value S using the formula (8) thr ；
2) Setting parameters, setting parameter alpha 1 、α 2 、β 1 、β 2 、β 3 The specific parameters are shown in table 2:
TABLE 2 automatic extraction method parameters of image of brain subcutaneous tissue patch
Wherein Tl is a distance coefficient between the maximum peak value of the whole image and the peak value of the target tissue, and the expression is as follows:
wherein fz A Numbering the maximum interval corresponding to the peak value in the gray distribution diagram of the whole image, fz T And numbering the peak values in the gray distribution diagram of the target tissue.
Step four, extracting a brain subcutaneous tissue patch image: the specific process comprises the steps of calculating membership by using the parameters through a formula (9), traversing a search space, searching a central region of the subcutaneous brain tissue with the maximum membership value, determining a final central position by combining an offset parameter delta y in a table 2, expanding by using the central position, and extracting a subcutaneous brain tissue patch image. In this embodiment, the extraction results of the 6 patches of the subcutaneous brain tissue are shown in fig. 3, and it can be seen from fig. 3 that the method successfully extracts the patch images of the subcutaneous brain tissue, and the subcutaneous brain tissue is located near the center of the patch images, which is beneficial to subsequent registration calculation.
Claims (4)
1. An automatic extraction method of a brain subcutaneous tissue patch image is characterized by comprising the following steps:
step one, setting the size of an initial search area and a central area of a brain subcutaneous tissue: analyzing and determining an initial search area of a brain subcutaneous tissue and the size of a brain subcutaneous tissue central area in a target image through a brain image data set;
step two, extracting an image gray distribution map: respectively extracting pixel gray distribution maps of a target image and a central region image of a brain subcutaneous tissue, wherein the specific method for extracting the image gray distribution maps comprises the following steps: carrying out pixel normalization processing on the image, dividing the pixel value of 0-1 into 20 intervals, counting the number of pixel points in each interval, and drawing a gray distribution graph by taking the pixel value as a horizontal coordinate and the number of the pixel points as a vertical coordinate;
step three, calculating the membership degree of the image of the brain subcutaneous tissue patch;
and step four, extracting the brain subcutaneous tissue patch image.
2. The method for automatically extracting the patch image of the subcutaneous brain tissue according to claim 1, wherein the first step comprises the specific process of setting the size of the initial search area and the size of the central area of the subcutaneous brain tissue as follows:
1) carrying out pixel value normalization processing on a gray level image in the brain image data set;
2) expanding all images in the brain image dataset to the same size by adopting a background pixel supplementing mode, and placing brain tissues in the images in the center of the images;
3) analyzing the boundary position of the brain subcutaneous tissue in the brain image data set, determining the left limit position, the right limit position, the upper limit position and the lower limit position, and setting a rectangular area as an initial search area of the brain subcutaneous tissue according to the four limit positions;
4) the gray distribution diagram of the brain subcutaneous tissue is approximately in normal distribution, the peak value in the gray distribution diagram is the characteristic gray value of the brain subcutaneous tissue, and the minimum rectangle size containing all the characteristic gray values is selected as the size of the central area of the brain subcutaneous tissue.
3. The method for automatically extracting the image of the brain subcutaneous tissue patch according to claim 1, wherein the step three, the specific process of calculating the membership degree of the image of the brain subcutaneous tissue patch is as follows:
1) meterCalculating pixel point proportion information D of peak value and neighborhood in brain subcutaneous tissue gray level distribution diagram s ：
Where i is the interval number corresponding to the peak in the distribution diagram, and N i Is the number of pixel points in i in the interval, n is the total number of pixel points in the central region, beta 1 ，β 2 ，β 3 Is a weight coefficient, wherein 1 Representing the relative position of the target peak to the peak of the entire image, beta 2 As peak ratio information weight, beta 2 Pixel site proportion weight of peak value and its neighborhood, N j The number of pixel points in the interval j;
2) calculating the similarity M of each interval gradient information in the gray distribution map of the target image and each interval gradient information in the gray distribution map of the central region of the subcutaneous brain tissue s ：
In the formula S thr As gradient threshold, Ng j Gradient values of adjacent intervals in the gray distribution;
S thr obtained by analysis of the brain image dataset, in the following way:
a. extracting the gray distribution map of the gray image in the brain image data set and the gray distribution map of the target image by adopting the method of the second step, and performing cubic B-spline optimization on the extracted gray distribution maps to avoid the interference of abnormal values;
b. calculating gradient value Ng of each interval j ；
Ng j ＝|N j+1 -N j |，j∈[0,19] (4)
NG＝[Ng 1 ,Ng 2 ,…,Ng 19 ] (5)
c. Fitting the gray distribution diagram of the gray image in the brain data set and the gradient variance of the gray distribution diagram of the target image, wherein the calculation formula is as follows:
in the formula, ANg j For gradient values of adjacent intervals in the gray profile of a gray image in a brain dataset, TNg j Is the gradient value of adjacent interval in the gray distribution of the target image, a j For the weight coefficient, tn is the number of maps, and S is calculated by the formulas (6), (7) and (8) thr ；
3) Calculating the membership Ey of the target tissue sub-image:
Ey＝α 1 M s +α 2 D s (9)
in the formula of alpha 1 And alpha 2 Are the weight coefficients.
4. The method for automatically extracting the patch image of the subcutaneous brain tissue according to claim 1, wherein the specific process of extracting the patch image of the subcutaneous brain tissue in the fourth step is as follows: traversing the search space, searching for a central region of the brain subcutaneous tissue with the maximum membership value, and extracting a brain subcutaneous tissue patch image according to the central region of the brain subcutaneous tissue.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210477260.2A CN114841958B (en) | 2022-05-04 | 2022-05-04 | Method for automatically extracting image of brain subcutaneous tissue patch |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210477260.2A CN114841958B (en) | 2022-05-04 | 2022-05-04 | Method for automatically extracting image of brain subcutaneous tissue patch |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114841958A true CN114841958A (en) | 2022-08-02 |
| CN114841958B CN114841958B (en) | 2023-02-07 |
Family
ID=82567628
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210477260.2A Active CN114841958B (en) | 2022-05-04 | 2022-05-04 | Method for automatically extracting image of brain subcutaneous tissue patch |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114841958B (en) |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102332054A (en) * | 2011-08-25 | 2012-01-25 | 西北工业大学 | Method for identifying cerebral cortex surface marking region of three-dimensional cerebral magnetic resonance image |
| US20120184843A1 (en) * | 2011-01-14 | 2012-07-19 | Yi-Hsuan Kao | Method and apparatus for brain perfusion magnetic resonance images |
| CN103310196A (en) * | 2013-06-13 | 2013-09-18 | 黑龙江大学 | Finger vein recognition method by interested areas and directional elements |
| CN106127749A (en) * | 2016-06-16 | 2016-11-16 | 华南理工大学 | The target part recognition methods of view-based access control model attention mechanism |
| CN108898602A (en) * | 2018-06-27 | 2018-11-27 | 南京邮电大学 | A kind of FCM medical image cutting method based on improvement QPSO |
| CN112164082A (en) * | 2020-10-09 | 2021-01-01 | 深圳市铱硙医疗科技有限公司 | Method for segmenting multi-modal MR brain image based on 3D convolutional neural network |
-
2022
- 2022-05-04 CN CN202210477260.2A patent/CN114841958B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20120184843A1 (en) * | 2011-01-14 | 2012-07-19 | Yi-Hsuan Kao | Method and apparatus for brain perfusion magnetic resonance images |
| CN102332054A (en) * | 2011-08-25 | 2012-01-25 | 西北工业大学 | Method for identifying cerebral cortex surface marking region of three-dimensional cerebral magnetic resonance image |
| CN103310196A (en) * | 2013-06-13 | 2013-09-18 | 黑龙江大学 | Finger vein recognition method by interested areas and directional elements |
| CN106127749A (en) * | 2016-06-16 | 2016-11-16 | 华南理工大学 | The target part recognition methods of view-based access control model attention mechanism |
| CN108898602A (en) * | 2018-06-27 | 2018-11-27 | 南京邮电大学 | A kind of FCM medical image cutting method based on improvement QPSO |
| CN112164082A (en) * | 2020-10-09 | 2021-01-01 | 深圳市铱硙医疗科技有限公司 | Method for segmenting multi-modal MR brain image based on 3D convolutional neural network |
Non-Patent Citations (1)
| Title |
|---|
| 周寿军等: "基于Markov随机场的脑部三维磁共振血管造影数据的分割", 《集成技术》 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN114841958B (en) | 2023-02-07 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Park et al. | Skull stripping based on region growing for magnetic resonance brain images | |
| Han et al. | CRUISE: cortical reconstruction using implicit surface evolution | |
| CN109753997B (en) | Automatic accurate robust segmentation method for liver tumor in CT image | |
| CN103942780B (en) | Based on the thalamus and its minor structure dividing method that improve fuzzy connectedness algorithm | |
| EP1597701A1 (en) | Method and apparatus for extracting cerebral ventricular system from images | |
| CN108986114B (en) | Automatic abdominal CT sequence image liver segmentation method based on level set and shape descriptor | |
| CN114359277B (en) | Brain image processing method and system for stroke patient | |
| SG176860A1 (en) | A method and system for segmenting a brain image | |
| Chen et al. | Shape-based automatic detection of pectoral muscle boundary in mammograms | |
| CN106778793A (en) | The repeatable measuring method and device of a kind of image feature | |
| CN106780492B (en) | Method for extracting key frame of CT pelvic image | |
| Tosun et al. | Cortical reconstruction using implicit surface evolution: accuracy and precision analysis | |
| US20150206300A1 (en) | Method and apparatus for extraction and quantification of hematoma from a brain scan such as computed tomography data | |
| Liu et al. | Automatic segmentation of the human brain ventricles from MR images by knowledge-based region growing and trimming | |
| CN110533667B (en) | Lung tumor CT image 3D segmentation method based on image pyramid fusion | |
| CN114841958B (en) | Method for automatically extracting image of brain subcutaneous tissue patch | |
| Krivov et al. | MRI augmentation via elastic registration for brain lesions segmentation | |
| CN112990367A (en) | Image processing method, device, equipment and storage medium | |
| Prima et al. | Automatic analysis of normal brain dissymmetry of males and females in MR images | |
| Hadjiprocopis et al. | Unbiased segmentation of diffusion-weighted magnetic resonance images of the brain using iterative clustering | |
| CN106875410B (en) | Automatic image contour extraction method and system | |
| Chansuparp et al. | The automated skull stripping of brain magnetic resonance images using the integrated method | |
| CN114913185A (en) | Texture segmentation method and system for lung CT image | |
| EP2535001A1 (en) | Method, a system and a computer program product for registration and identification of diagnostic images | |
| Abirami et al. | Automated brain tumor detection and identification using medical imaging |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |