[go: up one dir, main page]

CN113658208A - Cell image segmentation method of self-adaptive threshold - Google Patents

Cell image segmentation method of self-adaptive threshold Download PDF

Info

Publication number
CN113658208A
CN113658208A CN202110970486.1A CN202110970486A CN113658208A CN 113658208 A CN113658208 A CN 113658208A CN 202110970486 A CN202110970486 A CN 202110970486A CN 113658208 A CN113658208 A CN 113658208A
Authority
CN
China
Prior art keywords
image
area
segmentation
value
cell
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN202110970486.1A
Other languages
Chinese (zh)
Inventor
陈同生
刘曙通
孙晗
庄正飞
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Normal University Rayleigh Optoelectronic Technology Qingyuan Co ltd
Original Assignee
Normal University Rayleigh Optoelectronic Technology Qingyuan Co ltd
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Normal University Rayleigh Optoelectronic Technology Qingyuan Co ltd filed Critical Normal University Rayleigh Optoelectronic Technology Qingyuan Co ltd
Priority to CN202110970486.1A priority Critical patent/CN113658208A/en
Publication of CN113658208A publication Critical patent/CN113658208A/en
Pending legal-status Critical Current

Links

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
    • G06T7/136—Segmentation; Edge detection involving thresholding
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T5/00—Image enhancement or restoration
    • G06T5/20—Image enhancement or restoration using local operators
    • G06T5/30—Erosion or dilatation, e.g. thinning
    • 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
    • G06T7/155—Segmentation; Edge detection involving morphological operators
    • 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
    • G06T7/62—Analysis of geometric attributes of area, perimeter, diameter or volume
    • 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
    • G06T2207/10056—Microscopic image

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)

Abstract

The invention discloses a cell image segmentation method of a self-adaptive threshold, which comprises the following steps: acquiring a cell drug-adding processing image sequence comprising n images; after the grey value of the picture is normalized according to the maximum and minimum values, the grey value is mapped to a specified range in a linear mode [ m.n ]]In the method, the gray value of a pixel point corresponding to a picture is f (x, y); judging the fuzziness of the image, and mapping the image into a fuzzy domain; enhancing the effective information area, and inhibiting the ineffective information area, namely enhancing the fuzzy cell area and inhibiting the background area; selecting the enhanced images g 'of two adjacent frames'i，g′i‑1And (5) iteratively calculating an optimal segmentation threshold value of the image. The method realizes accurate and rapid segmentation of the cell delay image sequence, thereby combining the image characteristics extracted by the traditional method, inhibiting the problem of template over-expansion caused by area constraint and enhancing the segmentation effect of the image.

Description

Cell image segmentation method of self-adaptive threshold
Technical Field
The invention relates to the technical field of image processing, in particular to a cell image segmentation method capable of self-adapting to a threshold value.
Background
The dynamic analysis of living cells is of great significance for studying the structure, function and the regularity and essence of life activities of the cells.
In the dynamic quantitative analysis process, a cell region is often extracted from a time-delay sequence image of a cell for analysis and calculation, and the segmentation precision of the cell region also affects the final quantitative analysis result.
In recent years, segmentation algorithms for cell images have been developed rapidly, and conventional image segmentation methods are mainly classified into the following categories:
1) traditional threshold-based segmentation method
Threshold segmentation is a region segmentation technology, and is suitable for segmenting scenes with strong target-background contrast. The selection of the threshold is a key technology in the image threshold segmentation method. Among threshold-based image segmentation methods, the most typical representative is the maximum between-class variance threshold selection method (Otsus [1]) proposed by Otsu. For a cell image, when the target of the image is greatly different from the background, that is, the gray level histogram has no obvious double peaks, it is difficult to find an optimal segmentation threshold value based on the characteristics of a single image, the segmentation effect is not good, and some image preprocessing is often required to be added. However, for a sequence image with a large change in form and contrast, the preprocessing parameters need to be changed greatly, and each frame of image of the sequence image needs to be adjusted manually.
2) Traditional energy functional-based method
The segmentation method based on the energy functional mainly refers to an active contour model and an algorithm developed on the basis of the active contour model. The continuous curve serving as the independent variable is evolved to a target boundary by defining an energy functional, so that the segmentation process becomes a process for solving the minimum value of the function. The distance regularization level set algorithm (DRLSE [2]) is the well-known segmentation method among others. When the traditional energy functional algorithm is used for processing images with different forms and contrasts, the parameters of the evolution function need to be adjusted, so that the optimal segmentation effect is achieved. In the dynamic study of cells, specific drug treatment is often required to be added to study the reaction of the cells to the drugs, the morphology and contrast of the cells after being added with the drugs are greatly changed, and the energy functional algorithm parameters of each frame of image are required to be manually adjusted.
3) Method based on region growing
The segmentation method based on the edge detection obtains the boundary by utilizing the larger change of the gray value of the pixels on the edges of two adjacent regions, thereby realizing the image segmentation. The algorithm is sensitive to noise, holes are easily formed in the region, effective segmentation is difficult to achieve on a cell fluorescence image, and the segmentation time is long.
Disclosure of Invention
The invention aims to provide a cell image segmentation method with adaptive threshold, which realizes accurate and rapid segmentation of a cell delay image sequence and solves the problems in the background technology.
In order to achieve the purpose, the invention provides the following technical scheme:
a cell image segmentation method of adaptive threshold value comprises the following steps:
step 1: acquiring a cell drug-adding processing image sequence comprising n images;
step 2: after normalizing the gray value of the picture according to the maximum and minimum values, linearly mapping the gray value of the picture to a specified range [ m.n ], wherein the gray value of a pixel point corresponding to the picture is f (x, y), and the linear mapping formula is as follows:
Figure BDA0003225458370000021
and step 3: judging the fuzziness of the image, and mapping the image into a fuzzy domain, wherein the mapping formula is as follows:
Figure BDA0003225458370000022
wherein, gmaxIs the maximum value of the gray scale of the image after linear mapping, i.e. 255, g (i, j) is the pixel value of the image after linear mapping at the point (i, j);
and 4, step 4: the method is characterized in that the effective information area is enhanced, the ineffective information area is suppressed, namely, the fuzzy cell area is enhanced, the background area is suppressed, and the fuzzy enhancement formula is as follows:
Figure BDA0003225458370000023
and 5: selecting the enhanced images g 'of two adjacent frames'i，g′i-1Iteratively computing an optimal segmentation threshold th for the imagei＝βi·thbaseN-1 until the area constraint is satisfied:
Figure BDA0003225458370000031
Figure BDA0003225458370000032
i.e. the area S of the cell region divided by the current frameiCell region area S divided from the previous framei-1The ratio of the binary template to the binary template is near 1, and a binary template is generated;
step 6: in the backward frame-by-frame iteration process, after a certain number of frames of delta t are iterated each time, the shape of the current frame template area needs to be initialized, the excessive expansion or shrinkage of the template area caused by multiple iterations is prevented, the weighted sum of the segmentation area of the traditional segmentation algorithm and the segmentation area of the current beta threshold segmentation algorithm is used as an initialization area, and the initialization formula is as follows:
Sinitial＝μSo+λSa
wherein S isinitialIs the initialized area, S0Is the traditional binary coarse division area, SaIs the current beta threshold segmentation area, mu and lambda are weighting coefficients, satisfying the weighted sumIs 1.
Further, in step 2, during actual processing, the gray-scale values are mapped into a range from 0 to 255, that is, m is 255, n is 0, and the gray-scale values of the image are evenly distributed into the maximum range that the image can display through linear mapping, so that the contrast between the target area and the background area in the image is improved, and the gray-scale values of the target area and the background area are more different and better distinguished.
Further, the ambiguity function μ (i, j) e [0, 1] of step 3]The blur function takes 0.5 as a boundary, the region of the image gray value g (i, j) mapped to the blur field with a value μ (i, j) greater than 0.5 is taken as a valid information region, i.e. the target region to be emphasized, the region less than 0.5 is taken as an invalid information region, i.e. the background region, the positive parameter FdAnd FeThe image gray level at a blur level of 0.5 is controlled and is usually set to Fd＝10，FeWhen actually adjusted, F is 2eNormally, keep 2 constant and adjust F onlydThe value of (c).
Further, step 4 may embed and reuse a fuzzy enhancement formula to achieve a better enhancement effect, the number of times of enhancement is usually 2, after the enhancement is completed, inverse transformation is used to perform inverse transformation from a fuzzy domain to a gray value domain, the inverse transformation formula is:
Figure BDA0003225458370000033
furthermore, the lowerlimit and upperlimit in step 5 are the allowable cell area variation amounts of the adjacent frames, and usually, the lowerlimit is set to 0.002 and the upperlimit is set to 0.002, depending on the actual adjustment range.
Further, the optimal segmentation threshold th of step 6iIs selected to be equivalent to the optimum betaiSelection of the value, βiThe iterative formula of the value is betai＝βi+ Δ β, the initial value of the iteration is the optimal β value β of the previous framei-1The iteration step | Δ β | can be adjusted according to actual conditions, and is usually set to 0.01, and the iteration direction, i.e., the positive and negative values of Δ β, is determined by the following formula:
Figure BDA0003225458370000041
compared with the prior art, the invention has the beneficial effects that:
1) according to the invention, area constraints of adjacent frames are introduced, in an initial frame, an optimal segmentation multiple is manually selected to obtain an optimal segmentation cell region, morphological characteristics which are closest to reality of the cell region are retained, and then the morphological characteristics which are closer to reality are transmitted to the whole sequence through overlapping area constraints, so that each frame of image can obtain a better segmentation effect. The method is different from the traditional method for segmenting the image only based on certain morphological characteristics of the single-frame image, and introduces the integral morphological change characteristics of the cell sequence image and the real morphological characteristics of the cells introduced in the initial frame when segmenting the single-frame image through area constraint, thereby achieving the segmentation effect superior to the traditional method.
2) The invention also introduces the traditional image segmentation method to initialize the image segmentation template in a weighted sum mode, thereby combining the image characteristics extracted by the traditional method, inhibiting the problem of template over expansion caused by area constraint and enhancing the image segmentation effect.
3) The invention adopts a binarization threshold segmentation mode with the highest speed, and can obtain a higher segmentation speed compared with other traditional segmentation methods.
Drawings
FIG. 1 is a flow chart of a method of adaptive threshold cellular image segmentation in accordance with the present invention;
FIG. 2 is a contour overlay of the segmentation results of an adaptive threshold cellular image 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, a method for segmenting a cell image with adaptive threshold includes the following steps:
step 1: acquiring a cell drug-adding processing image sequence comprising n images;
step 2: after normalizing the gray value of the picture according to the maximum and minimum values, linearly mapping the gray value of the picture to a specified range [ m.n ], wherein the gray value of a pixel point corresponding to the picture is f (x, y), and the linear mapping formula is as follows:
Figure BDA0003225458370000051
in actual processing, the grayscale values are usually mapped to a range of 0 to 255, i.e., m is 255 and n is 0. Through linear mapping, the gray value of the image is evenly distributed in the maximum range which can be displayed by the image, so that the contrast between the target area (cell) and the background area in the image is improved, and the gray value difference between the target area (cell) and the background area is larger and better.
And step 3: judging the fuzziness of the image, and mapping the image into a fuzzy domain, wherein the mapping formula is as follows:
Figure BDA0003225458370000052
wherein, gmaxIs the maximum value of the gray scale of the image after linear mapping, i.e. 255, g (i, j) is the pixel value of the image after linear mapping at the point (i, j). Ambiguity function mu (i, j) is E [0,1 ∈]The blur function takes 0.5 as a boundary, the region of the image gray value g (i, j) mapped to the blur field with a value μ (i, j) greater than 0.5 is taken as a valid information region, i.e. the target region to be emphasized, the region less than 0.5 is taken as an invalid information region, i.e. the background region, the positive parameter FdAnd FeThe image gray level at a blur level of 0.5 is controlled and is usually set to Fd＝10，FeWhen actually adjusted, F is 2eUsually, 2 is kept constant and only adjustment is madeFdA value of (d);
and 4, step 4: the method is characterized in that the effective information area is enhanced, the ineffective information area is suppressed, namely, the fuzzy cell area is enhanced, the background area is suppressed, and the fuzzy enhancement formula is as follows:
Figure BDA0003225458370000053
the fuzzy enhancement formula can be embedded and reused to achieve a better enhancement effect, and the enhancement times are generally 2 times. After the enhancement is completed, the inverse transform is used to transform from the fuzzy domain to the grey value domain. The inverse transformation formula is:
Figure BDA0003225458370000061
and 5: selecting the enhanced images g 'of two adjacent frames'i，g′i-1Iteratively computing an optimal segmentation threshold th for the imagei＝βi·thbaseN-1 until the area constraint is satisfied:
Figure BDA0003225458370000062
Figure BDA0003225458370000063
i.e. the area S of the cell region divided by the current frameiCell region area S divided from the previous framei-1The ratio of the binary template to the binary template is near 1, and a binary template is generated; the lowerlimit and upperlimit are the cell area change amounts allowed for the adjacent frames, and usually, depending on the actual adjustment range, lowerlimit is set to 0.002 and upperlimit is set to 0.002. Wherein the optimal division multiple betaiIs a real number greater than 1, thbaseThe gray value corresponding to the first peak value of the frequency distribution histogram of the frame image. Optimal segmentation threshold thiIs selected to be equivalent to the optimum betaiAnd (4) selecting a value. Beta is aiThe iterative formula of the value is betai＝βi+ Δ β, the initial value of the iteration is the optimal β value β of the previous framei-1The iteration step | Δ β | can be adjusted according to actual conditions, and is usually set to 0.01, and the iteration direction, i.e., the positive and negative values of Δ β, is determined by the following formula:
Figure BDA0003225458370000064
optimal division multiple beta of start frame0Need to be manually set up, so that at th0＝β0·thbaseAnd under the threshold value, the segmentation effect is optimal.
Step 6: in the backward frame-by-frame iteration process, after a certain number of frames of delta t are iterated each time, the shape of the current frame template area needs to be initialized, the excessive expansion or shrinkage of the template area caused by multiple iterations is prevented, the weighted sum of the segmentation area of the traditional segmentation algorithm and the segmentation area of the current beta threshold segmentation algorithm is used as an initialization area, and the initialization formula is as follows:
Sinitial＝μSo+λSa
wherein S isinitialIs the initialized area, S0Is the traditional binary coarse division area, SaIs the current beta threshold segmentation area, mu and lambda are weighting coefficients, satisfying a weighted sum of 1.
The traditional initialization algorithm adopts an Otsu (maximum between-class variance) algorithm, and the image is set to contain L gray levels, and the number of pixels with gray values of i is set to be NiThe total number of pixels is:
N＝N0+N1+...+NL-1
the probability of a point with a gray value i is:
Figure BDA0003225458370000071
according to the desired formula, the mean value of the image gray levels is:
Figure BDA0003225458370000072
according to the figureGradation characteristics of the image, dividing the image into cell regions C by using a threshold value T0And background C1Two types, then ω0(T) and ω0(T) each represents C0And C1The probability of occurrence, i.e.:
Figure BDA0003225458370000073
ω1(T)＝1-ω0(T)
C0and C1The mean value of (A) is:
Figure BDA0003225458370000074
Figure BDA0003225458370000075
the inter-class variance with a threshold value T in the histogram is defined as:
Figure BDA0003225458370000076
when the inter-class variance is maximum, the segmentation threshold T is optimal:
Figure BDA0003225458370000077
comparing the performance of the Otsu algorithm and the DRLSE algorithm of the traditional image segmentation algorithm with the performance of the self-adaptive threshold algorithm provided by the invention, acquiring a Ca + dosing image sequence of 350 cells in an experiment, selecting images at the time point of 90 and the time point of 200 (shown in figure 2) at which the starting point is 0, the ending point is 350, and the morphology and the signal-to-noise ratio are obviously changed, respectively selecting three algorithms for segmentation, and evaluating by using three indexes of a cross-over ratio (IoU), a similarity coefficient (Dice) and an SN (sensitivity). The index is defined as follows:
1. cross-over ratio (IoU): and (5) performing intersection set ratio of the algorithm segmentation result and the standard result. IoU, the closer to 1, the better the coincidence of the standard result with the algorithm segmentation, and the higher the segmentation accuracy. The formula is as follows:
Figure BDA0003225458370000081
and 2, Dice (similarity coefficient) reflecting the similarity between the algorithm segmentation result and the standard result. A larger value indicates a higher accuracy of the segmentation result. The formula is as follows, wherein R is the algorithm segmentation result, and S is the standard segmentation result.
Figure BDA0003225458370000082
SN (sensitivity): the ratio of the number of correctly divided pixels to the number of pixels of the standard result. True Positive (TP) and False Negative (FN):
Figure BDA0003225458370000083
the results of the evaluation of the segmentation accuracy are shown in table 1:
Figure BDA0003225458370000084
Figure BDA0003225458370000091
results of the segmentation time are shown in Table 2
Figure BDA0003225458370000092
The above description is only for the preferred embodiment of the present invention, but the scope of the present invention is not limited thereto, and any person skilled in the art should be able to cover the technical solutions and the inventive concepts of the present invention within the technical scope of the present invention.

Claims (6)

1. A method for segmenting a cell image by self-adapting threshold value is characterized by comprising the following steps:
step 1: acquiring a cell drug-adding processing image sequence comprising n images;
step 2: after normalizing the gray value of the picture according to the maximum and minimum values, linearly mapping the gray value of the picture to a specified range [ m.n ], wherein the gray value of a pixel point corresponding to the picture is f (x, y), and the linear mapping formula is as follows:
Figure FDA0003225458360000011
and step 3: judging the fuzziness of the image, and mapping the image into a fuzzy domain, wherein the mapping formula is as follows:
Figure FDA0003225458360000012
wherein, gmaxIs the maximum value of the gray scale of the image after linear mapping, i.e. 255, g (i, j) is the pixel value of the image after linear mapping at the point (i, j);
and 4, step 4: the method is characterized in that the effective information area is enhanced, the ineffective information area is suppressed, namely, the fuzzy cell area is enhanced, the background area is suppressed, and the fuzzy enhancement formula is as follows:
Figure FDA0003225458360000013
and 5: selecting the enhanced images g 'of two adjacent frames'i，g′i-1Optimal segmentation threshold th of iterative computation imagei＝βi·thbaseN-1 until the area constraint is satisfied:
Figure FDA0003225458360000014
Figure FDA0003225458360000015
namely, the cell region area Si divided in the current frame and the cell region area S divided in the previous framei-1The ratio of the binary template to the binary template is near 1, and a binary template is generated;
step 6: in the backward frame-by-frame iteration process, after a certain number of frames of delta t are iterated each time, the shape of the current frame template area needs to be initialized, the excessive expansion or shrinkage of the template area caused by multiple iterations is prevented, the weighted sum of the segmentation area of the traditional segmentation algorithm and the segmentation area of the current beta threshold segmentation algorithm is used as an initialization area, and the initialization formula is as follows:
Sinitial＝μSo+λSa
wherein S isinitialIs the initialized area, S0Is the traditional binary coarse division area, SaIs the current beta threshold segmentation area, mu and lambda are weighting coefficients, satisfying a weighted sum of 1.
2. The adaptive threshold cellular image segmentation method according to claim 1, wherein in the step 2, during the actual processing, the gray-level values are mapped into a range from 0 to 255, i.e., m is 255 and n is 0, and the gray-level values of the image are evenly distributed into the maximum range that the image can display through linear mapping, so as to improve the contrast between the target region and the background region in the image, so that the gray-level values of the target region and the background region are more different and better distinguished.
3. An adaptive threshold cellular image segmentation method as defined in claim 1, wherein the ambiguity function μ (i, j) e [0, 1] of step 3]The blur function takes 0.5 as a boundary, the region of the image gray value g (i, j) mapped to the blur field with a value μ (i, j) greater than 0.5 is taken as a valid information region, i.e. the target region to be emphasized, the region less than 0.5 is taken as an invalid information region, i.e. the background region, the positive parameter FdAnd FeThe image gray level at a blur level of 0.5 is controlled and is usually set to Fd＝10，FeWhen actually adjusted, F is 2eNormally, keep 2 constant and adjust F onlydThe value of (c).
4. The adaptive threshold cellular image segmentation method of claim 1, wherein step 4 is performed by repeatedly using a fuzzy enhancement formula, usually 2 times, and performing inverse transformation from a fuzzy domain to a gray-scale domain after the enhancement is completed, wherein the inverse transformation formula is:
Figure FDA0003225458360000021
5. the adaptive threshold cell image segmentation method as claimed in claim 1, wherein the lowerlimit and upperlimit of step 5 are the allowable cell area variation of adjacent frames, and the lowerlimit is usually set to 0.002 and the upperlimit is set to 0.002 according to the actual adjustment range.
6. The adaptive threshold cellular image segmentation method according to claim 1, wherein the optimal segmentation threshold th of step 6iIs selected to be equivalent to the optimum betaiSelection of the value, βiThe iterative formula of the value is betai＝βi+ Δ β, the initial value of the iteration is the optimal β value β of the previous framei-1The iteration step | Δ β | can be adjusted according to actual conditions, and is usually set to 0.01, and the iteration direction, i.e., the positive and negative values of Δ β, is determined by the following formula:
Figure FDA0003225458360000022
CN202110970486.1A 2021-08-23 2021-08-23 Cell image segmentation method of self-adaptive threshold Pending CN113658208A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202110970486.1A CN113658208A (en) 2021-08-23 2021-08-23 Cell image segmentation method of self-adaptive threshold

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202110970486.1A CN113658208A (en) 2021-08-23 2021-08-23 Cell image segmentation method of self-adaptive threshold

Publications (1)

Publication Number Publication Date
CN113658208A true CN113658208A (en) 2021-11-16

Family

ID=78481682

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202110970486.1A Pending CN113658208A (en) 2021-08-23 2021-08-23 Cell image segmentation method of self-adaptive threshold

Country Status (1)

Country Link
CN (1) CN113658208A (en)

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
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
CN108682008A (en) * 2018-05-07 2018-10-19 江苏康尚生物医疗科技有限公司 A kind of Leukocyte Image clarity evaluation method and device
CN108961208A (en) * 2018-05-21 2018-12-07 江苏康尚生物医疗科技有限公司 A kind of aggregation leucocyte segmentation number system and method
CN110120056A (en) * 2019-05-21 2019-08-13 闽江学院 Blood leucocyte dividing method based on self-adapting histogram threshold value and contour detecting

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
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
CN108682008A (en) * 2018-05-07 2018-10-19 江苏康尚生物医疗科技有限公司 A kind of Leukocyte Image clarity evaluation method and device
CN108961208A (en) * 2018-05-21 2018-12-07 江苏康尚生物医疗科技有限公司 A kind of aggregation leucocyte segmentation number system and method
CN110120056A (en) * 2019-05-21 2019-08-13 闽江学院 Blood leucocyte dividing method based on self-adapting histogram threshold value and contour detecting

Similar Documents

Publication Publication Date Title
Li et al. SAR image change detection using PCANet guided by saliency detection
CN109766898B (en) Image character recognition method, device, computer equipment and storage medium
CN107564017B (en) Method for detecting and segmenting urban high-resolution remote sensing image shadow
CN116758059A (en) Visual nondestructive testing method for roadbed and pavement
CN111582198B (en) Automatic sea-land segmentation method for remote sensing image
CN105374015A (en) Binary method for low-quality document image based on local contract and estimation of stroke width
CN103679173A (en) Method for detecting image salient region
CN109712149B (en) Image segmentation method based on wavelet energy and fuzzy C-means
CN104537673A (en) Infrared image segmentation method based on multiple threshold values and self-adaptation fuzzy clustering
CN117372463B (en) An image segmentation optimization method for power component images
CN113658163A (en) High-score SAR image segmentation method based on multi-level collaboration to improve FCM
CN114897932B (en) Infrared target tracking realization method based on feature and gray level fusion
Zhao et al. A novel Neutrosophic image segmentation based on improved fuzzy C-means algorithm (NIS-IFCM)
CN111047559B (en) Method for rapidly detecting abnormal area of digital pathological section
CN106919950B (en) Brain MR Image Segmentation Method Based on Probability Density Weighted Geodesic Distance
CN103456017B (en) Image partition method based on the semi-supervised weight Kernel fuzzy clustering of subset
CN113392854A (en) Image texture feature extraction and classification method
CN113658208A (en) Cell image segmentation method of self-adaptive threshold
Akther et al. Detection of vehicle's number plate at nighttime using iterative threshold segmentation (ITS) algorithm
CN110853030B (en) Method for evaluating the quality of virus-infected cells in a bioreactor
CN117953215A (en) Infrared single small target detection method and system based on local difference and global feature
CN115082379B (en) A method for segmenting flocs and filamentous bacteria in activated sludge phase contrast microscopy images
CN116012323A (en) Image definition calculation method, image processing model training method and device
CN113920391B (en) Target counting method based on scale generation self-adaptive truth diagram
CN117152169A (en) Target segmentation method and device for depth image, terminal equipment and storage medium

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