CN111429461A - Novel segmentation method for overlapped exfoliated epithelial cells - Google Patents
Novel segmentation method for overlapped exfoliated epithelial cells Download PDFInfo
- Publication number
- CN111429461A CN111429461A CN201910019927.2A CN201910019927A CN111429461A CN 111429461 A CN111429461 A CN 111429461A CN 201910019927 A CN201910019927 A CN 201910019927A CN 111429461 A CN111429461 A CN 111429461A
- Authority
- CN
- China
- Prior art keywords
- image
- level set
- cell
- edge
- 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.)
- Granted
Links
- 230000011218 segmentation Effects 0.000 title claims abstract description 87
- 210000002919 epithelial cell Anatomy 0.000 title claims abstract description 74
- 238000000034 method Methods 0.000 title claims abstract description 64
- 238000010586 diagram Methods 0.000 claims abstract description 48
- 210000000805 cytoplasm Anatomy 0.000 claims abstract description 10
- 210000004027 cell Anatomy 0.000 claims description 76
- 210000003855 cell nucleus Anatomy 0.000 claims description 32
- 210000004940 nucleus Anatomy 0.000 claims description 25
- 230000008569 process Effects 0.000 claims description 22
- 230000006872 improvement Effects 0.000 claims description 13
- 238000003709 image segmentation Methods 0.000 claims description 8
- 238000011478 gradient descent method Methods 0.000 claims description 7
- 238000012545 processing Methods 0.000 claims description 7
- 230000002146 bilateral effect Effects 0.000 claims description 6
- 239000003086 colorant Substances 0.000 claims description 6
- 238000001914 filtration Methods 0.000 claims description 6
- 230000000877 morphologic effect Effects 0.000 claims description 5
- 206010040844 Skin exfoliation Diseases 0.000 claims description 4
- 230000035618 desquamation Effects 0.000 claims description 4
- 238000007781 pre-processing Methods 0.000 claims description 4
- 230000007797 corrosion Effects 0.000 claims description 3
- 238000005260 corrosion Methods 0.000 claims description 3
- 239000000284 extract Substances 0.000 claims description 2
- 239000008188 pellet Substances 0.000 claims 1
- 230000008859 change Effects 0.000 description 5
- 230000000630 rising effect Effects 0.000 description 4
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 4
- 238000001514 detection method Methods 0.000 description 3
- 238000003745 diagnosis Methods 0.000 description 3
- 206010028980 Neoplasm Diseases 0.000 description 2
- 230000002159 abnormal effect Effects 0.000 description 2
- 230000001154 acute effect Effects 0.000 description 2
- 238000004364 calculation method Methods 0.000 description 2
- 201000011510 cancer Diseases 0.000 description 2
- 230000000295 complement effect Effects 0.000 description 2
- 230000003902 lesion Effects 0.000 description 2
- 108090000623 proteins and genes Proteins 0.000 description 2
- 238000004458 analytical method Methods 0.000 description 1
- 230000002380 cytological effect Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 210000003608 fece Anatomy 0.000 description 1
- 238000007490 hematoxylin and eosin (H&E) staining Methods 0.000 description 1
- 206010020718 hyperplasia Diseases 0.000 description 1
- 210000002490 intestinal epithelial cell Anatomy 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000001575 pathological effect Effects 0.000 description 1
- 102000004169 proteins and genes Human genes 0.000 description 1
- 238000012216 screening Methods 0.000 description 1
- 238000010186 staining 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
- G06T5/00—Image enhancement or restoration
- G06T5/73—Deblurring; Sharpening
-
- 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
- G06T7/10—Segmentation; Edge detection
- G06T7/194—Segmentation; Edge detection involving foreground-background segmentation
-
- 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
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
- Y02A90/00—Technologies having an indirect contribution to adaptation to climate change
- Y02A90/10—Information and communication technologies [ICT] supporting adaptation to climate change, e.g. for weather forecasting or climate simulation
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Image Analysis (AREA)
- Image Processing (AREA)
Abstract
The invention discloses a novel segmentation method of overlapped exfoliated epithelial cells, which comprises a nucleus segmentation module and a cytoplasm segmentation module, wherein the nucleus segmentation module is implemented by combining an MSER algorithm and a Kmeans algorithm, and the cytoplasm segmentation integrates two segmentation methods of a Veno diagram and a level set.
Description
Technical Field
The invention relates to the technical field of medical image processing, in particular to a novel segmentation method of overlapped exfoliated epithelial cells.
Background
The exfoliated epithelial cells are a sample to be detected which can be conveniently collected and obtained clinically, and the screening diagnosis of abnormal hyperplasia or canceration cells in the exfoliated epithelial cells can be realized by examining or analyzing the exfoliated epithelial cells and utilizing novel biotechnology such as heteroploid cell morphological characteristic parameters, genetic material DNA content change and other gene chips in cell biology.
The exfoliated epithelial cells such as cervical exfoliated epithelial cells, oral exfoliated epithelial cells or intestinal epithelial cells exfoliated from feces are treated by a conventional cytological sample (such as Papanicolaou staining and HE staining), an experienced doctor is required to manually find a plurality of precancerous lesion cells from a large number of cells under a microscope, the working strength is high, people are easily tired, the operator is required to have high pathological knowledge and clinical experience, the diagnosis result is influenced by subjective factors of the operator and the like, and human errors are inevitable. Accurate and efficient early detection methods of cancer cells can help save more lives of patients with early cancer. In order to realize morphological analysis of precancerous lesion cells and automatic diagnosis of abnormal features, cell segmentation is the first step, so that the segmentation quality of the exfoliated epithelial cell image has a very important influence on the accuracy of a final detection result. The ideal cell image segmentation result not only reduces the complexity of subsequent classifier design, but also helps to improve the accuracy of final detection.
The method mainly comprises the steps of classifying pixels into different categories according to information of each pixel point in an image, and classifying the pixels into different categories according to a given judgment criterion.
Disclosure of Invention
The present invention is directed to a novel split epithelial cell overlap segmentation method to solve the problems set forth in the background art. In order to achieve the purpose, the invention provides the following technical scheme: a novel method for segmenting overlapping exfoliated epithelial cells.
In order to achieve the purpose, the invention provides the following technical scheme: a novel segmentation method of overlapped exfoliated epithelial cells comprises two modules of nucleus segmentation and cytoplasm segmentation;
the cell nucleus segmentation method integrates two segmentation algorithms of an MSER algorithm and a Kmeans algorithm, and the steps are as follows (assuming that an original exfoliated epithelial cell image to be segmented is imageOriginal):
firstly, graying and bilateral filtering preprocessing are carried out on imageOriginal, and a processing result is recorded as imagepreprocessing;
secondly, segmenting cell nucleuses in the image imagePreprocess by using an MSER algorithm to obtain a cell nucleus coarse segmentation result which is marked as nucleousRough;
generating an ROI picture for the rough segmentation contour of each cell nucleus in the image nucleolus, wherein the ROI picture is obtained by intercepting from the image imageProcess according to a rectangular region;
and step four, clustering the ROI picture by using a Kmeans algorithm to finally obtain an accurate segmentation result of the cell nucleus, and recording the accurate segmentation result as nucleousFine.
Furthermore, the exfoliated epithelial cell image imageOriginal is obtained by taking images under a microscope, the magnification of an ocular lens of the microscope is 10, and the magnification of an objective lens is 20.
Further, the MSER algorithm is developed based on a watershed algorithm, and the algorithm is used for segmenting the exfoliated epithelial cell nucleus by calculating the maximum stable extremum region in the image; the flow of the MSER algorithm is as follows: and (4) carrying out binarization on the image, and taking a binarization threshold value of [0,255], so that the binarized image undergoes a process from full black to full white, which is similar to an overhead view with the rising water level. In this process, the area of some connected domains has little change with the rising of the threshold, and such a region is called MSER, the maximum stable extremum region, and its mathematical definition is:
min_area＜Qi＜max_area
where Qi represents the area of the ith connected region and Δ represents a slight threshold change (water filling), when v (i) is less than a given threshold max _ variation and the area of the connected region is within a given minimum area threshold min _ area and maximum area threshold max _ area, then the region is considered to be the maximum stable extremum region that meets the requirements.
Further, the principle that the MSER algorithm locates exfoliated epithelial nuclei by calculating the maximum stable extremal region in the image is as follows: in the image of the exfoliated epithelial cell, the exfoliated epithelial cell nucleus is usually an elliptical region with a low gray value, which just marks the definition of the maximum stable extremum region in the MSER, so the exfoliated epithelial cell nucleus can be located by finding the maximum stable extremum region in the image of the exfoliated epithelial cell. Preferably, the value of the MSER algorithm delta is 2, the value of max _ variation is 0.5, the value of min _ area is 100, and the value of max _ area is 1500.
Further, the ROI picture is obtained by capturing from the image process according to a rectangular region, where the rectangular region is formed by diffusing L pixel points of the minimum circumscribed rectangle of the rough segmentation contour of the cell nucleus corresponding to the ROI picture, respectively, up, down, left, and right, and preferably L is equal to 16.
Further, the ROI picture is clustered by using a Kmeans algorithm, wherein the clustering number is 2 and corresponds to a nucleus region and a non-nucleus region in the ROI picture. According to the fact that the gray value of the exfoliated epithelial cell nucleus is usually lower than the surrounding cell pulp pixel points, the method extracts the area with the lower average gray value in the clustering result as the accurate area of the cell nucleus.
A cell paste segmentation method, wherein the cell paste segmentation method combines two segmentation methods of a Voronoi diagram and a level set, and comprises the following segmentation steps:
firstly, segmenting a desquamation epithelial cell mass by using a region growing method and combining a Kmeans algorithm, and marking the desquamation epithelial cell mass as ClumpImage.
From the coordinates of the center of mass of all nuclei in nucleosfine, a voronoi diagram of the image of exfoliated epithelial cells was generated, denoted voronoi image.
And step two, overlapping the images ClumpImage and Voronoi image to obtain the Voronoi diagram outlines of all the exfoliated epithelial cells, and recording the Voronoi diagram outlines as Voronoi segments.
And step three, taking the Voronoi diagram contour of the cell in the Voronoi segment image as an initial contour of the improved level set, and carrying out level set contour evolution so as to obtain an accurate contour cellFine of the cytoplasm.
Further, the exfoliated epithelial cell mass refers to a region composed of all cells in the exfoliated epithelial cell image, and the background region are mutually opposite and complementary, and the exfoliated epithelial cell mass is segmented, and the specific segmentation steps are as follows:
clustering image imageProcess subjected to graying and bilateral filtering pretreatment by using a Kmeans algorithm, wherein the clustering number is 2, and the clustering number is respectively corresponding to a background area and a cell cluster area in the image imageProcess to obtain a cell cluster rough segmentation result which is recorded as ClumpSegRough;
performing morphological corrosion operation on a background region in the ClumpSegRough image, so as to ensure that pixel points in the background region of the ClumpSegRough belong to a real background region to the maximum extent, and marking a processing result as ClumpSegRough;
and step three, randomly selecting a plurality of coordinate points in a background region of the ClumpSegRoughEversion, taking the coordinate points as seed points of a region growth algorithm, performing region growth on the image imageProcess by using the region growth algorithm, and finally obtaining the result of the region growth, namely the real region of the cell mass.
Further, the images ClumpImage and Voronoi image are mutually superposed to obtain the Voronoi image outlines of all the exfoliated epithelial cells, and the method is characterized in that the superposition mode is that RGB channel values of pixel points corresponding to the ClumpImage and the Voronoi image are added, and the remainder is taken for 255; the voronoi diagram profile is characterized in that the cytoplasm of different cells is represented by different colors, namely, connected domains of different colors in the diagram have one-to-one correspondence with the cytoplasm.
Furthermore, the level set belongs to an active contour theory, and the basic idea is to embed an evolution curve or an evolution curved surface into a high-dimensional level set function as a zero level set, and the purpose of controlling the evolution curve or the evolution curved surface is achieved through the high-dimensional level set function. The general steps for image segmentation using level sets are as follows:
providing an initial contour for a level set;
step two, constructing a level set function by using the initial contour as a zero level set;
thirdly, constructing an energy functional based on the level set function;
and step four, minimizing the energy functional through a gradient descent method, and realizing the evolution of a level set function, thereby realizing the evolution of a zero level set, wherein the evolution process of the zero level set is the process of the evolution of the target contour to be segmented.
And the level set minimizes the energy function through a gradient descent method, and the evolution of the level set function is realized, so that the evolution of the zero level set is realized, and the evolution process of the zero level set is the contour evolution process of the target to be segmented. The flow of the level set gradient descent algorithm is as follows:
input: initial Contour, gradient descent iteration number n.
Output: and 4, evolving the level set function after the end of the evolution.
Constructing an initial level set function phi by Contour
for itera in n:
returnΦ
Further, the improved level set is an improvement of the traditional DR L SE level set model aiming at the segmentation of overlapped exfoliated epithelial cells, and the improved level set mainly comprises the following improvements relative to the DR L SE model:
a. a more reasonable initial contour is provided for the level set by the voronoi diagram.
b. The edge enhancement is carried out on the edge indicator operator in the DR L SE, so that the interference of the edge of the adjacent overlapped cell on the contour evolution of the target to be segmented is effectively relieved.
c. Compared with the DR L SE edge energy item, the energy item can more accurately measure the degree of the zero level set at the edge position in the image, so that the improved level set can more fully and accurately utilize the edge information of the image to perform image segmentation.
The definition of the DR L SE edge energy term is as follows:
∫Ωg(Φ)|△Φ|dx
the edge energy term can be viewed as essentially integrating the edge indicator along a set of zero levels. When the edge of the target to be segmented is clear and the background of the target to be segmented is clean, the edge energy item can well measure the degree of the edge pixel point of the zero level set in the image. However, in fact, there are situations of edge blurring, crossing, breaking, etc. in the object to be segmented in the medical image that we process, and there are many interference edges in the background of the object to be segmented or in the object to be segmented itself, so there may be a situation that: most of the pixel points at the zero level are interference edge pixel points, but because the perimeter of the zero level set is very small, even if the zero level set does not really converge to the real edge of the target to be segmented, the energy item also obtains the minimum value, so that the zero level set is difficult to converge to the real edge of the target to be segmented.
The improvement of the edge energy item of the invention mainly aims at the situation that when the edge information in a medical image is too complex and the original edge energy item in DR L SE is used for segmenting the medical image, the perimeter of a zero level set in the edge energy item brings errors to image segmentation, therefore, the invention provides the following improved edge energy item:
the improved edge energy item is obtained by dividing the original edge energy item of DR L SE by the perimeter of the zero level set, so that the energy item can well measure the degree of the edge pixel point of the zero level set in the image.
Furthermore, the Voronoi diagram provides a more reasonable initial contour for the level set, the Voronoi diagram is generated by using the centroid coordinates of all the exfoliated epithelial cell nuclei, the Voronoi diagram segmentation result is usually closer to the real contour of the cell to be segmented, and the Voronoi diagram segmentation method has two advantages that the harsh requirement of a DR L SE model on the robustness of the initial contour is relieved on one hand, and on the other hand, the accurate contour of the object to be segmented can be evolved by the level set only with less contour evolution times, so that the efficiency of the level set is effectively improved.
Further, theThe edge enhancement is carried out on the edge indicating operator, mainly aiming at the condition of segmenting the overlapped and peeled epithelial cells, the purpose of the edge enhancement is to weaken the edge of the cell overlapped with the target cell to be segmented, and simultaneously retain the edge belonging to the target cell, thereby preventing the interference of the edge of the adjacent overlapped cell on the contour evolution of the target cell to the maximum extent. Edge enhancement for use with the present inventioniThe principle on which the method is based is that exfoliated epithelial cells are generally elliptical, the nucleus is generally located in the center of the cell, the angle formed by the vector formed by the connecting lines of the edge pixel points of the target nucleus and the target cell and the direction of the gradient at the pixel point should be an acute angle, and the angle formed by the vector formed by the connecting lines of the edge pixel points of the target cell and the adjacent overlapped cell and the direction of the gradient at the pixel point should be an obtuse angle.
DR L SE edge indicator:
the invention edge indication operator:
whereinαAnd representing an included angle formed by a vector formed by a pixel point in the edge indicator operator image and the cell nucleus centroid coordinate of the target cell to be segmented and the gradient direction of the pixel point.
Still further, the energy function of the improved level set of the present invention is defined as follows:
E(Φ)＝μRp(Φ)+λLf(Φ)+αAf(Φ)+E(Φ)，
wherein R isp(Φ) is the regular energy term, L f (Φ) is the edge energy term, Af (Φ) is the area energy term, E (Φ) is the shape prior energy term, and the definition of each energy sub-term is as follows:
the canonical energy term: rp(Φ)＝∫Ωp(|ΔΦ|)dx
Edge energy term:
area energy term: a. thef(Φ)＝∫ΩfH(-Φ)dx
Shape prior energy term: the invention considers that the prior shape of the cervical cell is an ellipse and names the prior term of the shape as an ellipse term, but the invention does not explicitly define the ellipse term, but directly gives the gradient descending flow of the ellipse term, and the calculation process of the gradient descending flow of the ellipse term is as follows: the minimum circumscribed ellipse epipse of the zero level set is fitted first, and then a new level set function phi _ epipse, i.e. the gradient descent flow of the ellipse term, is constructed using the epipse as a new zero level set.
The level set minimizes the energy function through a gradient descent method, so that the zero level set evolves to the real contour of the target to be segmented, and the gradient descent flow of the level set function is as follows:
aiming at the segmentation of overlapped exfoliated epithelial cells, the invention firstly provides the rough segmentation by using a Voronoi diagram and then carries out fine segmentation by using an improved level set, so that the Voronoi diagram segmentation result provides a more reasonable initial contour for the level set, thereby enabling the result of contour evolution of the level set to be more accurate, and on the other hand, the Voronoi diagram segmentation result is usually closer to the real contour of a cell, thereby enabling the level set to be optimal only by a few iterations, thereby reducing the time of contour evolution and improving the efficiency of cell segmentation.
Compared with the prior art, the invention has the following advantages and remarkable advantages:
a. the Voronoi diagram provides a reasonable initial contour for the level set by combining a two-stage segmentation process of the level set in an efficient and simple method, and the efficiency and the precision of contour evolution of the level set are improved;
b. the edge energy item of the traditional DR L SE level set is improved, and the improved level set is more suitable for segmentation of overlapped exfoliated epithelial cells;
c. the edge enhancement is carried out on the level set edge indicator, so that the interference of the edge of the adjacent overlapped cell on the profile evolution is relieved to the maximum extent;
d. an efficient cell mass segmentation algorithm based on Kmenas and region growth is provided.
Drawings
FIG. 1 is a general flow chart of a novel split epithelial cell segmentation method according to the present invention;
FIG. 2 is a flow chart of a cell mass segmentation method based on a region growing algorithm according to the present invention;
FIG. 3 is a schematic diagram of a nucleus segmentation method based on MSER algorithm of the present invention;
FIG. 4 is a diagram of the result of the nucleus segmentation based on the MSER algorithm of the present invention;
FIG. 5 is a schematic diagram of a cell mass segmentation method based on a region growing algorithm according to the present invention;
FIG. 6 is a schematic representation of the rough segmentation of exfoliated epithelial cells based on a Voronoi diagram in accordance with the present invention;
FIG. 7 is a graph showing the results of rough segmentation of exfoliated epithelial cells based on a voronoi diagram in accordance with the present invention;
FIG. 8 is a diagram of a pretreatment process for the novel split epithelial cell overlap segmentation method of the present invention;
FIG. 9 is a comparison of the level set edge indicator before and after improvement in accordance with the present invention;
FIG. 10 is a schematic diagram showing the contour evolution of a novel split epithelial cell segmentation method according to the present invention;
FIG. 11 is a graph showing the segmentation results of a novel split epithelial cell overlap segmentation method of the present invention;
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
Referring to fig. 1-11, the present invention provides the following technical solutions: the invention discloses a novel split epithelial cell overlapping segmentation method, which comprises a cell nucleus segmentation module and a cell pulp segmentation module, and the method combines an MSER algorithm and a Kmeans algorithm to segment a cell nucleus, and comprises the following steps (assuming that an original split epithelial cell image to be segmented is imageOriginal):
firstly, graying and bilateral filtering preprocessing are carried out on imageOriginal, and a processing result is recorded as imagepreprocessing;
secondly, segmenting cell nucleuses in the image imagePreprocess by using an MSER algorithm to obtain a cell nucleus coarse segmentation result which is marked as nucleousRough;
and step three, generating an ROI picture for the rough segmentation contour of each cell nucleus in the image nucleolus, wherein the ROI picture is obtained by intercepting the image imageProcess according to a rectangular region. Meanwhile, the exfoliated epithelial cell image imageOriginal is obtained by adopting images under the condition of 10 times of ocular magnification and 20 times of objective magnification of a microscope. The rectangular area is formed by diffusing 16 pixel points upwards, downwards, leftwards, rightwards and rightwards from the minimum external rectangle of the cell nucleus rough segmentation outline corresponding to the ROI picture;
and step four, clustering the ROI picture by using a Kmeans algorithm to finally obtain an accurate segmentation result of the cell nucleus, and recording the accurate segmentation result as nucleousFine.
The MSER algorithm is developed based on a watershed algorithm, and the MSER algorithm is used for segmenting exfoliated epithelial cell nuclei by calculating the maximum stable extremum region in an image; the flow of the MSER algorithm is as follows: and (4) carrying out binarization on the image, and taking a binarization threshold value of [0,255], so that the binarized image undergoes a process from full black to full white, which is similar to an overhead view with the rising water level. In this process, the area of some connected domains has little change with the rising of the threshold, and such a region is called MSER, the maximum stable extremum region, and its mathematical definition is:
min_area＜Qi＜max_area
where Qi represents the area of the ith connected region and Δ represents a slight threshold change (water filling), when v (i) is less than a given threshold max _ variation and the area of the connected region is within a given minimum area threshold min _ area and maximum area threshold max _ area, then the region is considered to be a satisfied MSER.
The principle of the MSER algorithm to locate exfoliated epithelial nuclei by calculating the most stable extremal region in the image is: in the image of the exfoliated epithelial cell, the exfoliated epithelial cell nucleus is usually an elliptical region with a low gray value, which just marks the definition of the maximum stable extremum region in the MSER, so the exfoliated epithelial cell nucleus can be located by finding the maximum stable extremum region in the image of the exfoliated epithelial cell. Meanwhile, the value of the MSER algorithm delta is 2, the value of max _ variation is 0.5, the value of min _ area is 100, and the value of max _ area is 1500.
The ROI pictures are clustered by using a Kmeans algorithm, the clustering number is 2, the clustering number corresponds to a cell nucleus region and a non-cell nucleus region, and because the gray value of a cast-off epithelial cell nucleus is usually lower than surrounding pixel points, a region with a lower average gray value in a clustering result is extracted as an accurate region of the cell nucleus.
A cell paste segmentation method, wherein the cell paste segmentation method combines two segmentation methods of a Voronoi diagram and a level set, and comprises the following segmentation steps:
a. firstly, segmenting a desquamation epithelial cell mass by using a region growth method and combining a Kmeans algorithm, and marking as ClumpImage;
b. generating a Voronoi image of the exfoliated epithelial cell image according to the coordinates of the mass centers of all cell nuclei in nucleosfine, and recording the Voronoi image;
c. superposing the images ClumpImage and Voronoi image, wherein the specific operation is to add RGB channel values of corresponding pixel points of the two images and take the remainder of 255 to obtain the Voronoi diagram outlines of all the exfoliated epithelial cells, and the Voronoi diagram outlines are marked as Voronoi segments; the cell pulp of different cells in the Voronoi diagram is represented by different colors, namely, connected domains with different colors in the diagram have one-to-one correspondence with the cell pulp;
d. carrying out level set contour evolution by taking the Voronoi diagram contour of the cell in the image Voronoi segment as an initial contour of an improved level set, thereby obtaining an accurate contour cellFine of cytoplasm;
the exfoliated epithelial cell mass refers to a region composed of all cells in an exfoliated epithelial cell image, and is mutually opposite and complementary with a background region, and the exfoliated epithelial cell mass is segmented, and the specific segmentation steps are as follows:
a. clustering the image process subjected to graying and bilateral filtering pretreatment by using a Kmeans algorithm, wherein the clustering number is 2, and the clustering number is respectively corresponding to a background region and a cell mass region in the exfoliated epithelial cell image to obtain a cell mass rough segmentation result which is marked as ClumpSegRough;
b. performing morphological corrosion operation on a background region in the ClumpSegRough image, thereby ensuring that pixel points in the roughly divided background region belong to a real background region to the maximum extent, and obtaining a ClumpSegRoughEjection processing result;
c. randomly selecting a plurality of coordinate points in a background region of the ClumpSegRoughEversion, taking the coordinate points as seed points of a region growing algorithm, performing region growing on the image imageProcess by using the region growing algorithm, and finally obtaining the result of the region growing, namely the real region of the cell mass.
The level set belongs to an active contour theory, and the basic idea is to embed an evolution curve or an evolution curved surface into a high-dimensional level set function as a zero level set, and the purpose of controlling the evolution curve or the evolution curved surface is achieved through the high-dimensional level set function. The general steps for image segmentation using level sets are as follows:
providing an initial contour for a level set;
step two, constructing a level set function by using the initial contour as a zero level set;
thirdly, constructing an energy functional based on the level set function;
and step four, minimizing the energy functional through a gradient descent method, and realizing the evolution of a level set function, thereby realizing the evolution of a zero level set, wherein the evolution process of the zero level set is the process of the evolution of the target contour to be segmented.
And the level set minimizes the energy function through a gradient descent method, and the evolution of the level set function is realized, so that the evolution of the zero level set is realized, and the evolution process of the zero level set is the contour evolution process of the target to be segmented. The flow of the level set gradient descent algorithm is as follows:
input: initial Contour, gradient descent iteration number n.
Output: and 4, evolving the level set function after the end of the evolution.
Constructing an initial level set function phi by Contour
for itera in n:
returnΦ
The improved level set is an improvement of the traditional DR L SE level set model aiming at the segmentation of overlapped exfoliated epithelial cells, and mainly comprises the following three improvements relative to the DR L SE model:
a. a more reasonable initial contour is provided for the level set by the voronoi diagram.
b. The edge enhancement is carried out on the edge indicator operator in the DR L SE, so that the interference of the edge of the adjacent overlapped cell on the contour evolution of the target to be segmented is effectively relieved.
c. Compared with the DR L SE edge energy item, the energy item can more accurately measure the degree of the zero level set at the edge position in the image, so that the improved level set can more fully and accurately utilize the edge information of the image to perform image segmentation.
The definition of the DR L SE edge energy term is as follows:
∫Ωg(Φ)|△Φ|dx
the edge energy term can be viewed as essentially integrating the edge indicator along a set of zero levels. When the edge of the target to be segmented is clear and the background of the target to be segmented is clean, the edge energy item can well measure the degree of the edge pixel point of the zero level set in the image. However, in fact, there are situations of edge blurring, crossing, breaking, etc. in the object to be segmented in the medical image that we process, and there are many interference edges in the background of the object to be segmented or in the object to be segmented itself, so there may be a situation that: most of the pixel points at the zero level are interference edge pixel points, but because the perimeter of the zero level set is very small, even if the zero level set does not really converge to the real edge of the target to be segmented, the energy item also obtains the minimum value, so that the zero level set is difficult to converge to the real edge of the target to be segmented.
The improvement of the edge energy item of the invention mainly aims at the situation that when the edge information in a medical image is too complex and the original edge energy item in DR L SE is used for segmenting the medical image, the perimeter of a zero level set in the edge energy item brings errors to image segmentation, therefore, the invention provides the following improved edge energy item:
the improved edge energy item is obtained by dividing the original edge energy item of DR L SE by the perimeter of the zero level set, so that the energy item can well measure the degree of the edge pixel point of the zero level set in the image.
The Voronoi diagram provides a more reasonable initial contour for the level set, the Voronoi diagram is generated by using the centroid coordinates of all the exfoliated epithelial cell nuclei, the Voronoi diagram segmentation result is usually closer to the real contour of the cell to be segmented, and the Voronoi diagram segmentation method has two advantages that the harsh requirement of a DR L SE model on the robustness of the initial contour is relieved on one hand, and on the other hand, the accurate contour of the object to be segmented can be evolved by the level set only by a few contour evolution times, so that the efficiency of the level set is effectively improved.
The edge enhancement is mainly aimed at the condition of segmenting the overlapped and exfoliated epithelial cells, the edge enhancement aims to weaken the edge of the cell overlapped with the target cell to be segmented, and simultaneously reserve the edge belonging to the target cell, so that the interference of the edge of the adjacent overlapped cell on the contour evolution of the target cell is prevented to the maximum extent. Edge enhancement for use with the present inventioniiThe principle on which the method is based is that exfoliated epithelial cells are generally elliptical, the nucleus is generally located in the center of the cell, the angle formed by the vector formed by the connecting lines of the edge pixel points of the target nucleus and the target cell and the direction of the gradient at the pixel point should be an acute angle, and the angle formed by the vector formed by the connecting lines of the edge pixel points of the target cell and the adjacent overlapped cell and the direction of the gradient at the pixel point should be an obtuse angle.
DR L SE edge indicator:
the invention edge indication operator:
whereinαAnd representing an included angle formed by a vector formed by a pixel point in the edge indicator operator image and the cell nucleus centroid coordinate of the target cell to be segmented and the gradient direction of the pixel point.
The more reasonable edge energy term is characterized in that aiming at the problem that a zero level set is difficult to converge to a target edge when a DR L SE edge energy term is used for processing a complex edge situation, the invention provides the following more reasonable edge energy term which is defined as follows:
the improvement of the energy item is that the perimeter of the zero level set is divided by the original edge energy item of DR L SE, so that the energy item can better measure the degree of the zero level set located at the edge pixel point in the image.
The energy function of the improved level set of the present invention is defined as follows:
E(Φ)＝μRp(Φ)+λLf(Φ)+αAf(Φ)+EE(Φ)，
wherein R isp(Φ) is the regular energy term, L f (Φ) is the edge energy term, Af (Φ) is the area energy term, E (Φ) is the shape prior energy term, and the definition of each energy sub-term is as follows:
the canonical energy term: rp(Φ)=∫Ωp(|ΔΦ|)dx
Edge energy term:
area energy term: a. thef(Φ)＝∫ΩfH(-Φ)dx
Shape prior energy term: the invention considers that the prior shape of the cervical cell is an ellipse and names the prior term of the shape as an ellipse term, but the invention does not explicitly define the ellipse term, but directly gives the gradient descending flow of the ellipse term, and the calculation process of the gradient descending flow of the ellipse term is as follows: the minimum circumscribed ellipse epipse of the zero level set is fitted first, and then a new level set function phi _ epipse, i.e. the gradient descent flow of the ellipse term, is constructed using the epipse as a new zero level set.
The level set minimizes the energy function through a gradient descent method, so that the zero level set evolves to the real contour of the target to be segmented, and the gradient descent flow of the level set function is as follows:
aiming at the segmentation of overlapped exfoliated epithelial cells, the invention firstly provides the rough segmentation by using a Voronoi diagram and then carries out fine segmentation by using an improved level set, so that the Voronoi diagram segmentation result provides a more reasonable initial contour for the level set, thereby enabling the result of contour evolution of the level set to be more accurate, and on the other hand, the Voronoi diagram segmentation result is usually closer to the real contour of a cell, thereby enabling the level set to be optimal only by a few iterations, thereby reducing the time of contour evolution and improving the efficiency of cell segmentation.
Finally, it should be noted that: although the present invention has been described in detail with reference to the foregoing embodiments, it will be apparent to those skilled in the art that changes may be made in the embodiments and/or equivalents thereof without departing from the spirit and scope of the invention. Any modification, equivalent replacement, or improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.
Claims (12)
1. A novel method for segmenting overlapping exfoliated epithelial cells is characterized in that: the method comprises two modules of cell nucleus segmentation and cell pulp segmentation.
2. The segmentation of nuclei of claim 1, characterized by: the invention uses a method combining MSER algorithm and Kmeans algorithm to segment the cell nucleus, and the steps are as follows (assuming that the original exfoliated epithelial cell image to be segmented is imageOriginal):
firstly, graying and bilateral filtering preprocessing are carried out on imageOriginal, and the result is recorded as imagepreprocessing;
secondly, segmenting cell nucleuses in the image imagePreprocess by using an MSER algorithm to obtain a cell nucleus coarse segmentation result which is marked as nucleousRough;
generating an ROI picture for the rough segmentation outline of each cell nucleus in the image nucleolus rough, wherein the ROI picture is obtained by intercepting from the image imageProcess according to a rectangular region, and the rectangular region is formed by diffusing L pixel points to the upper part, the lower part, the left part and the right part of the minimum circumscribed rectangle of the rough segmentation outline of the cell nucleus corresponding to the ROI picture respectively;
and step four, clustering the ROI picture by using a Kmeans algorithm to finally obtain the accurate contour of the cell nucleus, and marking as nucleosFine.
3. The exfoliated epithelial cell image imageOriginal as set forth in claim 2, wherein: the image was taken under a microscope with an eyepiece magnification of 10 and an objective magnification of 20.
4. The use of the Kmeans algorithm for clustering ROI pictures as claimed in claim 2, wherein: the Kmeans clustering number is 2, corresponding to a cell nucleus region and a non-cell nucleus region, and because the gray value of the exfoliated epithelial cell nucleus is usually lower than that of surrounding pixel points, the method extracts a region with a lower average gray value in a clustering result as an accurate region of the cell nucleus.
5. The cell paste dividing method according to claim 1, wherein: the cytoplasm segmentation fuses two segmentation methods of a voronoi diagram and a level set, and the segmentation steps are as follows:
firstly, segmenting a desquamation epithelial cell mass by using a region growth method and combining a Kmeans algorithm, and marking as ClumpImage;
generating a Voronoi image of the exfoliated epithelial cell image according to the mass center coordinates of all cell nuclei in the image nucleosfine, and recording the Voronoi image;
step three, overlapping the images ClumpImage and Voronoi image to obtain the Voronoi diagram outlines of all the exfoliated epithelial cells, and recording the Voronoi diagram outlines as Voronoi segments;
step four, taking the Voronoi diagram contour of the cell in the Voronoi segment image as the initial contour of the improved level set, and carrying out level set contour evolution to obtain the accurate contour cellFine of the cell pulp;
6. the cell pellet segmentation as claimed in step one of claim 5, wherein: the process of dividing the cell mass can be divided into the following steps:
clustering image processes subjected to graying and bilateral filtering preprocessing by using a Kmeans algorithm, wherein the clustering number is 2, the clustering number is respectively corresponding to a background region and a cell mass region in an exfoliated epithelial cell image, and a cell mass rough segmentation result is obtained and is marked as ClumpSegRough;
performing morphological corrosion treatment on a background region in the ClumpSegRough image, so as to ensure that pixel points in the ClumpSegRough background region belong to a real background region to the maximum extent, and marking a treatment result as ClumpSegRough;
and step three, randomly selecting a plurality of coordinate points in a background region of the ClumpSegRoughEversion, taking the coordinate points as seed points of a region growth algorithm, performing region growth on the image imageProcess by using the region growth algorithm, and finally obtaining the result of the region growth, namely the accurate region of the cell mass.
7. The method for obtaining the Voronoi image contour of the exfoliated epithelial cell according to the claim 5, wherein the images ClumpImage and Voronoi image are mutually superposed to obtain the Voronoi image contour of all the exfoliated epithelial cells, and the mutual superposition mode is characterized in that RGB channel values of pixel points corresponding to the ClumpImage and the Voronoi image are added, and 255 is left; the voronoi diagram profile is characterized in that the cytoplasm of different cells in the diagram is represented by different colors, namely, connected domains of different colors in the diagram have one-to-one correspondence with the cytoplasm.
8. The improved level set according to step four of claim 5, wherein the improved level set is an improvement of the conventional DR L SE level set model aiming at the segmentation of overlapping exfoliated epithelial cells, and the improved level set mainly comprises the following three improvements compared with the DR L SE model:
a. a more reasonable initial contour is provided for the level set through the voronoi diagram;
b. edge enhancement is carried out on an edge indicator operator in DR L SE, so that the interference of the edge of adjacent overlapped cells on the contour evolution of a target to be segmented is effectively relieved;
c. compared with the DR L SE edge energy item, the energy item can more accurately measure the degree of the zero level set at the edge position in the image, so that the improved level set can more fully and accurately utilize the edge information of the image to perform image segmentation.
9. The voronoi diagram described in claim 8 refinement a provides a more reasonable initial profile for the level set, characterized in that: the present invention uses the coordinates of the center of mass of all exfoliated epithelial nuclei to generate a voronoi diagram, the results of which segmentation have typically been compared to the true contour of the cells to be segmented.
10. The improvement of claim 8 wherein the edge enhancement of the edge indicator is performed by comparing the edge enhancement of the present invention with the DR L SE original edge indicator as follows:
DR L SE edge indicator:
the invention edge indication operator:
whereinαAnd representing an included angle formed by a vector formed by a pixel point in the edge indicator operator image and the cell nucleus centroid coordinate of the target cell to be segmented and the gradient direction of the pixel point.
11. The invention discloses a more reasonable edge energy term as claimed in claim 8, which is characterized in that, aiming at the problem that the DR L SE edge energy term has zero level set to be difficult to converge to the target edge when processing complex edge situation, the invention provides the following more reasonable edge energy term, which is defined as follows:
the improvement of the energy item is that the perimeter of the zero level set is divided by the original edge energy item of DR L SE, so that the energy item can better measure the degree of the zero level set located at the edge pixel point in the image.
12. The improved level set according to claim 5, step four, characterized in that its energy function is as follows:
E(Φ)＝μRp(Φ)+λLf(Φ)+αAf(Φ)+E(Φ)，
where Rp (Φ) is the canonical energy term, Lf(Φ) is the marginal energy term, Af(Φ) is the area energy term, E (Φ) is the shape prior energy term, and the definition of each energy sub-term is as follows:
the canonical energy term: rp(Φ)＝∫Ωp(|ΔΦ|)dx
Edge energy term:
area energy term: a. thef(Φ)＝∫ΩfH(-Φ)dx
The level set minimizes the energy function through a gradient descent method, so that the zero level set evolves to the real contour of the target to be segmented, and the gradient descent flow of the level set function is as follows:
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910019927.2A CN111429461B (en) | 2019-01-09 | 2019-01-09 | Novel segmentation method for overlapped and exfoliated epithelial cells |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910019927.2A CN111429461B (en) | 2019-01-09 | 2019-01-09 | Novel segmentation method for overlapped and exfoliated epithelial cells |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111429461A true CN111429461A (en) | 2020-07-17 |
| CN111429461B CN111429461B (en) | 2023-09-01 |
Family
ID=71545695
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910019927.2A Active CN111429461B (en) | 2019-01-09 | 2019-01-09 | Novel segmentation method for overlapped and exfoliated epithelial cells |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111429461B (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112053355A (en) * | 2020-09-16 | 2020-12-08 | 昆明理工大学 | Cell image segmentation method |
| CN114596563A (en) * | 2021-12-15 | 2022-06-07 | 江苏康尚生物医疗科技有限公司 | A method and system for self-adaptive segmentation and identification of tiny particles with strong anti-interference ability |
| CN116580041A (en) * | 2023-05-30 | 2023-08-11 | 山东第一医科大学附属眼科研究所(山东省眼科研究所、山东第一医科大学附属青岛眼科医院) | Corneal endothelial cell boundary segmentation method and device based on voronoi diagram |
| WO2024119326A1 (en) * | 2022-12-05 | 2024-06-13 | 深圳华大生命科学研究院 | Gene image data correction method and system, electronic device and storage medium |
Citations (19)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101238487A (en) * | 2005-06-15 | 2008-08-06 | 组织诺斯替斯有限责任公司 | Method for the segmentation of leukocytes |
| US20090136103A1 (en) * | 2005-06-24 | 2009-05-28 | Milan Sonka | System and methods for image segmentation in N-dimensional space |
| US20100040276A1 (en) * | 2007-03-23 | 2010-02-18 | Fraunhofer-Gesellschaft Zur Foerderung Der Angewan Dten Forschung E.V. | Method and apparatus for determining a cell contour of a cell |
| TW201124945A (en) * | 2010-01-06 | 2011-07-16 | Chunghwa Telecom Co Ltd | Method and system for automatically detecting representative region of pseudo-object-based image. |
| CN102831607A (en) * | 2012-08-08 | 2012-12-19 | 深圳市迈科龙生物技术有限公司 | Method for segmenting cervix uteri liquid base cell image |
| CN102831608A (en) * | 2012-08-06 | 2012-12-19 | 哈尔滨工业大学 | Unsteady measurement algorithm based image segmentation method of improved rule distance level set |
| CN103345748A (en) * | 2013-06-26 | 2013-10-09 | 福建师范大学 | Positioning and partition method for human tissue cell two-photon microscopic image |
| US20140023270A1 (en) * | 2012-07-19 | 2014-01-23 | Qualcomm Incorporated | Parameter Selection and Coarse Localization of Interest Regions for MSER Processing |
| CN104376559A (en) * | 2014-11-14 | 2015-02-25 | 中国人民解放军军事医学科学院卫生装备研究所 | Medical image segmentation method based on improved range adjustment level set algorithm |
| US20150078648A1 (en) * | 2013-09-13 | 2015-03-19 | National Cheng Kung University | Cell image segmentation method and a nuclear-to-cytoplasmic ratio evaluation method using the same |
| US20150169985A1 (en) * | 2013-12-12 | 2015-06-18 | University Of Pittsburgh - Of The Commonwealth System Of Higher Education | Automated epithelial nuclei segmentation for computational disease detection algorithms |
| CN104992435A (en) * | 2015-06-24 | 2015-10-21 | 广西师范大学 | Cervix uteri single cell image segmentation algorithm |
| US20160098589A1 (en) * | 2014-08-29 | 2016-04-07 | Definiens Ag | Applying Pixelwise Descriptors to a Target Image that are Generated by Segmenting Objects in Other Images |
| CN105574528A (en) * | 2015-12-15 | 2016-05-11 | 安徽工业大学 | Synechia cell image segmenting method based on polyphase mutual exclusion level set |
| CN105976367A (en) * | 2016-04-29 | 2016-09-28 | 上海联影医疗科技有限公司 | Image segmentation method, pulmonary nodule detection method and computer-aided detection system |
| US20170124709A1 (en) * | 2014-05-14 | 2017-05-04 | Massachusetts Institute Of Technology | Systems and methods for medical image segmentation and analysis |
| CN106650731A (en) * | 2016-12-23 | 2017-05-10 | 中山大学 | Robust license plate and logo recognition method |
| CN107730498A (en) * | 2017-10-30 | 2018-02-23 | 天津理工大学 | Novel uterine neck cell core partitioning algorithm |
| CN108596932A (en) * | 2018-04-18 | 2018-09-28 | 哈尔滨理工大学 | A Segmentation Method for Overlapped Cervical Cell Images |
-
2019
- 2019-01-09 CN CN201910019927.2A patent/CN111429461B/en active Active
Patent Citations (19)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101238487A (en) * | 2005-06-15 | 2008-08-06 | 组织诺斯替斯有限责任公司 | Method for the segmentation of leukocytes |
| US20090136103A1 (en) * | 2005-06-24 | 2009-05-28 | Milan Sonka | System and methods for image segmentation in N-dimensional space |
| US20100040276A1 (en) * | 2007-03-23 | 2010-02-18 | Fraunhofer-Gesellschaft Zur Foerderung Der Angewan Dten Forschung E.V. | Method and apparatus for determining a cell contour of a cell |
| TW201124945A (en) * | 2010-01-06 | 2011-07-16 | Chunghwa Telecom Co Ltd | Method and system for automatically detecting representative region of pseudo-object-based image. |
| US20140023270A1 (en) * | 2012-07-19 | 2014-01-23 | Qualcomm Incorporated | Parameter Selection and Coarse Localization of Interest Regions for MSER Processing |
| CN102831608A (en) * | 2012-08-06 | 2012-12-19 | 哈尔滨工业大学 | Unsteady measurement algorithm based image segmentation method of improved rule distance level set |
| CN102831607A (en) * | 2012-08-08 | 2012-12-19 | 深圳市迈科龙生物技术有限公司 | Method for segmenting cervix uteri liquid base cell image |
| CN103345748A (en) * | 2013-06-26 | 2013-10-09 | 福建师范大学 | Positioning and partition method for human tissue cell two-photon microscopic image |
| US20150078648A1 (en) * | 2013-09-13 | 2015-03-19 | National Cheng Kung University | Cell image segmentation method and a nuclear-to-cytoplasmic ratio evaluation method using the same |
| US20150169985A1 (en) * | 2013-12-12 | 2015-06-18 | University Of Pittsburgh - Of The Commonwealth System Of Higher Education | Automated epithelial nuclei segmentation for computational disease detection algorithms |
| US20170124709A1 (en) * | 2014-05-14 | 2017-05-04 | Massachusetts Institute Of Technology | Systems and methods for medical image segmentation and analysis |
| US20160098589A1 (en) * | 2014-08-29 | 2016-04-07 | Definiens Ag | Applying Pixelwise Descriptors to a Target Image that are Generated by Segmenting Objects in Other Images |
| CN104376559A (en) * | 2014-11-14 | 2015-02-25 | 中国人民解放军军事医学科学院卫生装备研究所 | Medical image segmentation method based on improved range adjustment level set algorithm |
| CN104992435A (en) * | 2015-06-24 | 2015-10-21 | 广西师范大学 | Cervix uteri single cell image segmentation algorithm |
| CN105574528A (en) * | 2015-12-15 | 2016-05-11 | 安徽工业大学 | Synechia cell image segmenting method based on polyphase mutual exclusion level set |
| CN105976367A (en) * | 2016-04-29 | 2016-09-28 | 上海联影医疗科技有限公司 | Image segmentation method, pulmonary nodule detection method and computer-aided detection system |
| CN106650731A (en) * | 2016-12-23 | 2017-05-10 | 中山大学 | Robust license plate and logo recognition method |
| CN107730498A (en) * | 2017-10-30 | 2018-02-23 | 天津理工大学 | Novel uterine neck cell core partitioning algorithm |
| CN108596932A (en) * | 2018-04-18 | 2018-09-28 | 哈尔滨理工大学 | A Segmentation Method for Overlapped Cervical Cell Images |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112053355A (en) * | 2020-09-16 | 2020-12-08 | 昆明理工大学 | Cell image segmentation method |
| CN114596563A (en) * | 2021-12-15 | 2022-06-07 | 江苏康尚生物医疗科技有限公司 | A method and system for self-adaptive segmentation and identification of tiny particles with strong anti-interference ability |
| WO2024119326A1 (en) * | 2022-12-05 | 2024-06-13 | 深圳华大生命科学研究院 | Gene image data correction method and system, electronic device and storage medium |
| CN116580041A (en) * | 2023-05-30 | 2023-08-11 | 山东第一医科大学附属眼科研究所(山东省眼科研究所、山东第一医科大学附属青岛眼科医院) | Corneal endothelial cell boundary segmentation method and device based on voronoi diagram |
| CN116580041B (en) * | 2023-05-30 | 2024-06-18 | 山东第一医科大学附属眼科研究所(山东省眼科研究所、山东第一医科大学附属青岛眼科医院) | Corneal endothelial cell boundary segmentation method and device based on voronoi diagram |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111429461B (en) | 2023-09-01 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN113723573B (en) | Tumor tissue pathological classification system and method based on adaptive proportion learning | |
| Song et al. | A deep learning based framework for accurate segmentation of cervical cytoplasm and nuclei | |
| CN104392460B (en) | A Segmentation Method of Adhesive Leukocytes Based on Nuclear Marker Watershed Transform | |
| CN103984958B (en) | Cervical cancer cell dividing method and system | |
| CN112215800B (en) | Overlapping Chromosome Identification and Segmentation Method Based on Machine Learning | |
| CN104484877B (en) | A kind of AML cell segmentation methods clustered based on Meanshift with morphological operation | |
| CN111429461A (en) | Novel segmentation method for overlapped exfoliated epithelial cells | |
| CN102214355B (en) | Cutout method of clothing display material | |
| CN107256558A (en) | The cervical cell image automatic segmentation method and system of a kind of unsupervised formula | |
| CN107492088B (en) | Automatic identification and statistics method for white blood cells in gynecological microscopic image | |
| CN106096491B (en) | An automated method for identifying microaneurysms in fundus color photographic images | |
| CN113570628B (en) | White blood cell segmentation method based on movable contour model | |
| CN104599271A (en) | CIE Lab color space based gray threshold segmentation method | |
| CN103489187A (en) | Quality test based segmenting method of cell nucleuses in cervical LCT image | |
| CN110889846A (en) | Diabetes retina image optic disk segmentation method based on FCM | |
| CN109033936A (en) | A kind of cervical exfoliated cell core image-recognizing method | |
| CN108320289A (en) | A kind of bone marrow cell dividing method based on rarefaction representation and morphological operation | |
| Soni et al. | CT scan based brain tumor recognition and extraction using Prewitt and morphological dilation | |
| CN115187852A (en) | Tibetan medicine urine diagnosis suspended matter identification method and device | |
| CN113850792A (en) | Cell classification counting method and system based on computer vision | |
| CN110210578B (en) | Cervical cancer histopathology microscopic image clustering system based on graph theory | |
| CN110458042B (en) | Method for detecting number of probes in fluorescent CTC | |
| CN108629780B (en) | Tongue image segmentation method based on color decomposition and threshold technology | |
| CN116385435B (en) | Pharmaceutical capsule counting method based on image segmentation | |
| CN115082379B (en) | A method for segmenting flocs and filamentous bacteria in activated sludge phase contrast microscopy images |
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 | ||
| CB02 | Change of applicant information | ||
| CB02 | Change of applicant information |
Address after: Floor 1 and 2, unit B, C and D, building B7, medical instrument Park, 818 Gaoxin Avenue, Donghu New Technology Development Zone, Wuhan City, Hubei Province, 430000 Applicant after: Wuhan Lanting intelligent Medicine Co.,Ltd. Address before: 430073 floor 1 and 2, unit B, C and D, building B7, medical instrument Park, 818 Gaoxin Avenue, Donghu New Technology Development Zone, Hongshan District, Wuhan City, Hubei Province Applicant before: WUHAN LANDING MEDICAL HI-TECH Ltd. |
|
| GR01 | Patent grant | ||
| GR01 | Patent grant |