CN112381838A - Automatic image cutting method for digital pathological section image - Google Patents
Automatic image cutting method for digital pathological section image Download PDFInfo
- Publication number
- CN112381838A CN112381838A CN202011273550.2A CN202011273550A CN112381838A CN 112381838 A CN112381838 A CN 112381838A CN 202011273550 A CN202011273550 A CN 202011273550A CN 112381838 A CN112381838 A CN 112381838A
- Authority
- CN
- China
- Prior art keywords
- image
- coordinate
- sparse matrix
- elements
- image block
- 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
- 238000005520 cutting process Methods 0.000 title claims abstract description 30
- 238000000034 method Methods 0.000 title claims abstract description 11
- 230000001575 pathological effect Effects 0.000 title claims abstract description 7
- 239000011159 matrix material Substances 0.000 claims abstract description 43
- 230000007170 pathology Effects 0.000 claims description 5
- 150000001875 compounds Chemical class 0.000 claims description 3
- 238000013135 deep learning Methods 0.000 claims description 3
- 238000013507 mapping Methods 0.000 claims description 3
- 238000005070 sampling Methods 0.000 abstract description 6
- 238000010586 diagram Methods 0.000 abstract 1
- 238000013461 design Methods 0.000 description 4
- 238000004364 calculation method Methods 0.000 description 3
- 238000001514 detection method Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000009826 distribution Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 238000009827 uniform distribution Methods 0.000 description 1
Images
Classifications
-
- 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
- G06T3/00—Geometric image transformations in the plane of the image
- G06T3/40—Scaling of whole images or parts thereof, e.g. expanding or contracting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/187—Segmentation; Edge detection involving region growing; involving region merging; involving connected component labelling
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/60—Analysis of geometric attributes
- G06T7/66—Analysis of geometric attributes of image moments or centre of gravity
-
- 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
- G06T2207/20—Special algorithmic details
- G06T2207/20021—Dividing image into blocks, subimages or windows
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Geometry (AREA)
- Image Analysis (AREA)
- Apparatus For Radiation Diagnosis (AREA)
Abstract
The invention discloses an automatic image cutting method of a digital pathological section image, which comprises the following steps: s1, recognizing and extracting the outline of the tissue area; s2, calculating the centroid coordinate and the length of the long axis of the contour; s3, setting parameters of the image block to be extracted; s4, calculating an intelligent graph cutting factor R; s5, designing a sparse matrix A; s6, calculating a coordinate matrix XY of a non-zero element of A; s7, calculating a coordinate matrix XY' in the 1x diagram; s8, whether the elements in the traversal XY' are inside the outline or not is judged, and whether the image block T is cut or not is judged. The invention has high sampling rate, ensures the accuracy of the cutting chart and has higher efficiency.
Description
Technical Field
The invention relates to the technical field of computer vision correlation, in particular to an automatic image cutting method for a digital pathological section image.
Background
The digital pathology full-section image is mainly a 40x image, the length and width distribution of one digital pathology full-section image is from tens of thousands of pixel units to hundreds of thousands of pixel units, and if the large image of the level needs to be analyzed by an algorithm, the image of a tissue region in the large image needs to be cut out for processing.
The image cutting and extracting method in the prior art has large calculation amount, and cannot give consideration to the image cutting efficiency while ensuring the accuracy.
Disclosure of Invention
The invention aims to provide an automatic image cutting method for a digital pathological section image, which has high sampling rate, ensures the accuracy of image cutting and has high efficiency.
In order to achieve the purpose, the invention is realized by adopting the following technical scheme:
the invention discloses an automatic image cutting method of a digital pathological section image, which comprises the following steps:
s1, reducing the digital pathology full-slice image I from a 40x image to a 1x image I ', extracting a tissue region mask of the 1x image I' through a deep learning FCN network, and generating a contour M;
s2, calculating the centroid coordinate (x) of the contour Mc,yc) And the length of the major axis L;
s3, setting parameters of the image block to be extracted, comprising the following steps:
s31, traversing each image block T in the image I to obtain the ith image block TCenter coordinate (x)i,yi),
S32, mapping each image block T to the 1x image I 'to obtain the center coordinate (x) of the image block T' on the 1x image I1,y1) Wherein
S33, calculating the center coordinate (x) of the image block T1,y1) With coordinates of center of mass (x)c,yc) A distance dist of wherein
S4, calculating an intelligent graph cutting factor R with the formula of
Wherein f () represents rounding down, max [ x, y ] represents taking the maximum of x and y;
s5, calculating a sparse matrix A, wherein the size of the sparse matrix A corresponds to the image block T, the row elements and the column elements of the sparse matrix A are h and w, R nonzero elements are arranged in the row elements of the sparse matrix A, R nonzero elements are arranged in the column elements of the sparse matrix A, and the nonzero elements in the sparse matrix A are uniformly distributed;
s6, obtaining a coordinate matrix XY of the non-zero elements in the sparse matrix A,
when R is not equal to n, the compound is,
wherein x0=iw,x1=iw+(w-1)/(n-1),…,xn=iw+n*(w-1)/(n-1),
y0=ih,y1=ih+(h-1)/(n-1),…,yn=ih+n*(h-1)/(n-1);
S7, obtaining a coordinate matrix XY 'of the sparse matrix A in which the non-zero elements are mapped to the 1x figure I',
s8, whether the traversal coordinate matrix XY' is within the contour M,
if the coordinate is in the outline M, the image block is subjected to image cutting operation,
if the coordinates are not within the contour M, then the operation is discarded.
Preferably, in step S5, R ═ 2,
preferably, in step S5, R ═ 3,
preferably, in step S6, R ═ 3,
wherein x0=iw,x1=iw+(w-1)/2,x2=iw+(w-1),
y0=ih,y1=ih+(h-1)/2,y2=ih+(h-1)。
The invention has the beneficial effects that:
1. the calculated amount of the whole algorithm is controlled by the intelligent graph cutting factor R in a small amount of calculation to achieve the purpose of judgment, and the efficiency is high.
2. The intelligent image cutting factor R adopts the design idea of detecting the ratio of the distance between the central coordinate of the image block T and the centroid of the tissue outline to the length of the long axis of the outline, and has high accuracy.
3. The invention uses the non-zero element uniform distribution in the sparse matrix A, and improves the sampling efficiency.
Drawings
FIG. 1 is a schematic flow chart of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is further described in detail below with reference to the accompanying drawings.
As shown in fig. 1, the present invention comprises the steps of:
s1, reducing the digital pathology full-slice image I from a 40x image to a 1x image I ', extracting a tissue region mask of the 1x image I' through a deep learning FCN network, and generating a contour M;
s2, calculating the centroid coordinate (x) of the contour Mc,yc) And the length of the major axis L;
s3, setting parameters of the image block to be extracted, comprising the following steps:
s31, traversing each image block T in the image I, obtaining the center coordinate (x) of the ith image block Ti,yi),
S32, mapping each image block T to the 1x image I 'to obtain the center coordinate (x) of the image block T' on the 1x image I1,y1) Wherein
S33, calculating the center coordinate (x) of the image block T1,y1) With coordinates of center of mass (x)c,yc) A distance dist of wherein
S4, calculating an intelligent graph cutting factor R with the formula of
Wherein f () represents rounding down, max [ x, y ] represents taking the maximum value of x and y, and the value range of R is [2, 10 ];
s5, calculating a sparse matrix A, wherein the size of the sparse matrix A corresponds to the image block T, the row elements and the column elements of the sparse matrix A are h and w, R nonzero elements are arranged in the row elements of the sparse matrix A, R nonzero elements are arranged in the column elements of the sparse matrix A, and the nonzero elements in the sparse matrix A are uniformly distributed;
s6, obtaining a coordinate matrix XY of the non-zero elements in the sparse matrix A,
when R is not equal to n, the compound is,
wherein x0=iw,x1=iw+(w-1)/(n-1),…,xn=iw+n*(w-1)/(n-1),
y0=ih,y1=ih+(h-1)/(n-1),…,yn=ih+n*(h-1)/(n-1);
S7, obtaining a coordinate matrix XY 'of the sparse matrix A in which the non-zero elements are mapped to the 1x figure I',
s8, whether the traversal coordinate matrix XY' is within the contour M,
if the coordinate is in the outline M, the image block is subjected to image cutting operation,
if the coordinates are not within the contour M, then the operation is discarded.
For example, in step S5, when R is 2,
when R is 3 in step S5,
in step S6, for example, R ═ 3,
wherein x0=iw,x1=iw+(w-1)/2,x2=iw+(w-1),
y0=ih,y1=ih+(h-1)/2,y2=ih+(h-1)。
In the invention, the value of the intelligent map cutting factor R is [2, 10]]The number of the non-zero elements of the sparse matrix A is R2The size of the coordinate matrix XY of the non-zero elements of A is R, and whether the number of points in the contour is R needs to be judged2. There are w x h points in block T, and only R is discriminated by using the algorithm2The point can judge whether the block T and the tissue area have intersection, thereby greatly reducing the calculation amount.
Design idea of intelligent graph cutting factor R: the ratio of the distance between the central coordinate of the detection block T and the centroid of the tissue profile to the length of the long axis of the profile is larger, and the larger the L/dist is, the closer the detection block T is to the tissue profile is, and the density of sampling points needs to be improved. The larger the L/dist is, the larger the calculated R value is, the more the number of the non-zero elements of the corresponding sparse matrix A is, the density of the sampling points is improved, and the accuracy of the judgment of the tangent graph is ensured.
The value range [2, 10] of R is limited, thus providing guarantee for reducing the calculated amount.
Design idea of sparse matrix A:
1. the larger the value of the intelligent graph cutting factor R is, the corresponding number R of the non-zero elements of the sparse matrix A is2The more the cutting patterns are, the more the cutting patterns are connected with the design idea of the intelligent cutting pattern factor R, and the accuracy of cutting pattern judgment is ensured.
2. Non-zero elements in the sparse matrix A are uniformly distributed, so that the sampling efficiency is improved, the R value can be effectively sampled in a lower range, and the calculated amount is reduced.
The present invention is capable of other embodiments, and various changes and modifications may be made by one skilled in the art without departing from the spirit and scope of the invention.
Claims (4)
1. An automatic image cutting method for digital pathological section images is characterized by comprising the following steps:
s1, reducing the digital pathology full-slice image I from a 40x image to a 1x image I ', extracting a tissue region mask of the 1x image I' through a deep learning FCN network, and generating a contour M;
s2, calculating the centroid coordinate (x) of the contour Mc,yc) And the length of the major axis L;
s3, setting parameters of the image block to be extracted, comprising the following steps:
s31, traversing each image block T in the image I, obtaining the center coordinate (x) of the ith image block Ti,yi),
S32, mapping each image block T to the 1x image I 'to obtain the center coordinate (x) of the image block T' on the 1x image I1,y1) Wherein
S33, calculating the center coordinate (x) of the image block T1,y1) With coordinates of center of mass (x)c,yc) A distance dist of wherein
S4, calculating an intelligent graph cutting factor R with the formula of
Wherein f () represents rounding down, max [ x, y ] represents taking the maximum of x and y;
s5, calculating a sparse matrix A, wherein the size of the sparse matrix A corresponds to the image block T, the row elements and the column elements of the sparse matrix A are h and w, R nonzero elements are arranged in the row elements of the sparse matrix A, R nonzero elements are arranged in the column elements of the sparse matrix A, and the nonzero elements in the sparse matrix A are uniformly distributed;
s6, obtaining a coordinate matrix XY of the non-zero elements in the sparse matrix A,
when R is not equal to n, the compound is,
wherein x0=iw,x1=iw+(w-1)/(n-1),...,xn=iw+n*(w-1)/(n-1),
y0=ih,y1=ih+(h-1)/(n-1),...,yn=ih+n*(h-1)/(n-1);
S7, obtaining a coordinate matrix XY 'of the sparse matrix A in which the non-zero elements are mapped to the 1x figure I',
s8, whether the traversal coordinate matrix XY' is within the contour M,
if the coordinate is in the outline M, the image block is subjected to image cutting operation,
if the coordinates are not within the contour M, then the operation is discarded.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202011273550.2A CN112381838B (en) | 2020-11-14 | 2020-11-14 | Automatic image cutting method for digital pathological section image |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202011273550.2A CN112381838B (en) | 2020-11-14 | 2020-11-14 | Automatic image cutting method for digital pathological section image |
Publications (2)
Publication Number | Publication Date |
---|---|
CN112381838A true CN112381838A (en) | 2021-02-19 |
CN112381838B CN112381838B (en) | 2022-04-19 |
Family
ID=74582460
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202011273550.2A Active CN112381838B (en) | 2020-11-14 | 2020-11-14 | Automatic image cutting method for digital pathological section image |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN112381838B (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN113505766A (en) * | 2021-09-09 | 2021-10-15 | 北京智源人工智能研究院 | Image target detection method and device, electronic equipment and storage medium |
Citations (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106204625A (en) * | 2016-07-27 | 2016-12-07 | 大连理工大学 | A kind of variable focal length flexibility pose vision measuring method |
CN108470176A (en) * | 2018-01-24 | 2018-08-31 | 浙江科技学院 | A kind of notable extracting method of stereo-picture vision indicated based on frequency-domain sparse |
CN109002747A (en) * | 2017-06-07 | 2018-12-14 | 北京航空航天大学 | Visible remote sensing image Ship Detection on star based on geometrical characteristic |
CN109118473A (en) * | 2018-07-03 | 2019-01-01 | 深圳大学 | Angular-point detection method, storage medium and image processing system neural network based |
US20190003979A1 (en) * | 2015-12-09 | 2019-01-03 | Indevr, Inc. | Automated agglutination analyzer with contour comparison |
CN110033454A (en) * | 2019-04-19 | 2019-07-19 | 河北大学 | The dividing method for the lung neoplasm that large area adhesion lung boundary is organized in CT image |
CN110825123A (en) * | 2019-10-21 | 2020-02-21 | 哈尔滨理工大学 | Control system and method for automatic following loading vehicle based on motion algorithm |
CN111080658A (en) * | 2019-12-16 | 2020-04-28 | 中南民族大学 | Cervical MRI image segmentation method based on deformable registration and DCNN |
CN111292315A (en) * | 2020-03-05 | 2020-06-16 | 四川大学华西医院 | Rapid registration algorithm for pathological section tissue area |
-
2020
- 2020-11-14 CN CN202011273550.2A patent/CN112381838B/en active Active
Patent Citations (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20190003979A1 (en) * | 2015-12-09 | 2019-01-03 | Indevr, Inc. | Automated agglutination analyzer with contour comparison |
CN106204625A (en) * | 2016-07-27 | 2016-12-07 | 大连理工大学 | A kind of variable focal length flexibility pose vision measuring method |
CN109002747A (en) * | 2017-06-07 | 2018-12-14 | 北京航空航天大学 | Visible remote sensing image Ship Detection on star based on geometrical characteristic |
CN108470176A (en) * | 2018-01-24 | 2018-08-31 | 浙江科技学院 | A kind of notable extracting method of stereo-picture vision indicated based on frequency-domain sparse |
CN109118473A (en) * | 2018-07-03 | 2019-01-01 | 深圳大学 | Angular-point detection method, storage medium and image processing system neural network based |
CN110033454A (en) * | 2019-04-19 | 2019-07-19 | 河北大学 | The dividing method for the lung neoplasm that large area adhesion lung boundary is organized in CT image |
CN110825123A (en) * | 2019-10-21 | 2020-02-21 | 哈尔滨理工大学 | Control system and method for automatic following loading vehicle based on motion algorithm |
CN111080658A (en) * | 2019-12-16 | 2020-04-28 | 中南民族大学 | Cervical MRI image segmentation method based on deformable registration and DCNN |
CN111292315A (en) * | 2020-03-05 | 2020-06-16 | 四川大学华西医院 | Rapid registration algorithm for pathological section tissue area |
Non-Patent Citations (5)
Title |
---|
PIETRO,CERVERI等: "Automated method for computing the morphological and clinical parameters of the proximal femur using heuristic modeling techniques", 《ANNALS OF BIOMEDICAL ENGINEERING》 * |
SAMUELWILLIAMS等: "Optimization of sparse matrix-vector multiplication on emerging multicore platforms", 《PARALLEL COMPUTING》 * |
张小山: "基于遥感图像和AIS数据的综合船舶识别***设计与实现", 《中国优秀硕士学位论文全文数据库 (工程科技Ⅱ辑)》 * |
曹传东等: "基于质心距离倒数积分的轮廓匹配方法研究", 《中国航空学会》 * |
甄勇等: "高分辨率SAR图像舰船目标几何结构特征提取", 《信号处理》 * |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN113505766A (en) * | 2021-09-09 | 2021-10-15 | 北京智源人工智能研究院 | Image target detection method and device, electronic equipment and storage medium |
CN113505766B (en) * | 2021-09-09 | 2022-01-04 | 北京智源人工智能研究院 | Image target detection method and device, electronic equipment and storage medium |
Also Published As
Publication number | Publication date |
---|---|
CN112381838B (en) | 2022-04-19 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN109903319B (en) | Multi-resolution-based fast iteration closest point registration algorithm | |
CN113409267B (en) | Pavement crack detection and segmentation method based on deep learning | |
CN109272467A (en) | A kind of stratification image partition method based on multi-scale edge clue | |
CN110473174B (en) | Method for calculating accurate number of pencils based on image | |
CN112381838B (en) | Automatic image cutting method for digital pathological section image | |
CN109191493B (en) | Target tracking method based on RefineNet neural network and sparse optical flow | |
CN115578615A (en) | Night traffic sign image detection model establishing method based on deep learning | |
CN113177897A (en) | Rapid lossless filtering method for disordered 3D point cloud | |
CN112711981A (en) | High-precision transmission line defect identification method | |
CN106324708B (en) | Digitizing solution, the device of rainfall record drawing | |
CN108846845B (en) | SAR image segmentation method based on thumbnail and hierarchical fuzzy clustering | |
CN114140466A (en) | Plant root system measuring method, system and device based on image processing | |
CN110781943A (en) | Clustering method based on adjacent grid search | |
CN117495735A (en) | Automatic building elevation texture repairing method and system based on structure guidance | |
CN112164065B (en) | Real-time image semantic segmentation method based on lightweight convolutional neural network | |
CN108830878A (en) | A kind of method for tracking target based on FPN neural network | |
CN113487555A (en) | Point cloud gridding-based power transmission line hidden danger point rapid detection method | |
CN111404075B (en) | Automatic power transmission line measurement method for unmanned aerial vehicle inspection | |
CN108510010A (en) | A kind of density peaks clustering method and system based on prescreening | |
CN106251004B (en) | The Target cluster dividing method divided based on room for improvement distance | |
CN110009035B (en) | Wind measuring station cluster space clustering method based on image matching | |
CN112258536A (en) | Integrated positioning and dividing method for corpus callosum and lumbricus cerebellum | |
CN109739942B (en) | Saddle point extraction method based on contour line model | |
CN115100613A (en) | Lane line detection method based on laser radar point cloud information | |
CN111723737B (en) | Target detection method based on multi-scale matching strategy deep feature learning |
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 |