CN103020488A - Subcellular localization method based on fluorescent microscopic image - Google Patents
Subcellular localization method based on fluorescent microscopic image Download PDFInfo
- Publication number
- CN103020488A CN103020488A CN2012105881105A CN201210588110A CN103020488A CN 103020488 A CN103020488 A CN 103020488A CN 2012105881105 A CN2012105881105 A CN 2012105881105A CN 201210588110 A CN201210588110 A CN 201210588110A CN 103020488 A CN103020488 A CN 103020488A
- Authority
- CN
- China
- Prior art keywords
- image
- pixel
- subcellular localization
- subcellular
- value
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Images
Abstract
The invention discloses a subcellular localization method based on a fluorescent microscopic image. The subcellular localization method comprises the steps as follows: firstly, calculating the average value mu of all pixels with pixel values of more than 30 for each subcellular localization image, assigning zero to the pixels with the pixel values of less than mu-30, and obtaining the pre-processed subcellular localization image; secondly, calculating difference values of each pixel point with a grey value of more than zero and eight adjacent points in the image, marking the adjacent points with the pixel values of more than those of central pixel points as 1, and otherwise, marking as 0, wherein the number of the 8-domain values of 1 in each central pixel point is 0-8; thirdly, dividing the pixel points into 9 types according to the number of the adjacent points marked as 1 around the pixel points; fourthly, counting the number of each type of the pixel points in the image; fifthly, obtaining the sum of positive values, the sum of negative values and the sum of absolute values for the adjacent points of each type of the pixel points according to a difference matrix model of the image pixel points, and obtaining 36 characteristic quantities in total; and lastly, training and predicting the image by utilizing a support vector machine algorithm according to the extracted 36 groups of subcellular image characteristic quantities.
Description
Technical field
The invention belongs to the biology information technology field, particularly a kind of Subcellular Localization method based on fluorescence microscope images.
Background technology
Epr gene has led the rapid growth of gene and protein sequencing, and now, people's concern power has turned to the function of coded protein.Subcellular Localization refers to that certain albumen or expression product are at intracellular concrete Present site.Such as only existence in nuclear, or exist in the kytoplasm, or exist on the cell membrane, this has very important significance in bioinformatics.The at present automatically development of fluorescence microscopy so that Protein Subcellular can be in high-throughout situation the imaging location, therefore need fast automatically computing technique and Effective arithmetic come to the subcellular fraction image effectively quantize, identification and classification.
At present, people are applied to field of biology to computer technology, utilize computing machine to replace the part mankind's brainwork that the biological information of magnanimity is carried out identification and classification, therefrom draw valuable information [2].Image statistics is verified to have very successful application aspect the differentiation location, but its method commonly used calculating is relatively slow, and often need to select separately cell from experimental image, thereby has limited the scope of processing power and potential application.
Summary of the invention
The purpose of this invention is to provide a kind of Subcellular Localization method based on fluorescence microscope images.
Technical scheme of the present invention is that a kind of Subcellular Localization method based on fluorescence microscope images may further comprise the steps:
Step 1, at first the Subcellular Localization image based on fluorescence microscope images is carried out pre-service, to each secondary Subcellular Localization image, calculate its pixel value greater than the average u of all pixels of 30, then value being composed less than the pixel value of the pixel of u-30 is zero, obtain pretreated Subcellular Localization image
Step 3, extract the Subcellular Localization characteristics of image, namely by step 2, at first count the pixel number that this each class of Subcellular Localization image comprises, secondly, according to the matrix of differences model of Subcellular Localization image slices vegetarian refreshments, add up each class Subcellular Localization image slices vegetarian refreshments greater than zero value summation and minus value summation, again according to the absolute value sum of each pixel difference in each class, can obtain the variance of each class at last, so altogether obtain 36 stack features amounts;
Step 4 according to 36 groups of Subcellular Localization image feature amount extracting, is utilized algorithm of support vector machine, and the Subcellular Localization image is trained prediction, realizes the Subcellular Localization based on fluorescence microscope images.
The Subcellular Localization image carried out pre-treatment step comprise described in the step 1,
At first, for the Subcellular Localization image, filter out subcellular fraction image intermediate value greater than 30 pixel value, and calculate its average u, as shown in Equation 1:
Sum?Value=∑(I>30).（1）
u=mean(Sum?Value).
Then this image intermediate value being composed less than the pixel value of u-30 is zero, as shown in Equation 2,
I=0,if(I<u-30)（2）
The step that subcellular fraction framing pixel described in the step 2 is adjacent difference a little comprises,
According to the Subcellular Localization image that obtains after the pre-service, its pixel value scope is 0, u-30～255, in image I〉0 pixel, calculate the difference of itself and abutment points, as shown in Equation 3, difference greater than zero be labeled as 1, other be labeled as 0, obtain 9 kinds of model matrixs in the following table, a kind of situation in the corresponding table of each pixel obtains (0)-(8) 9 kinds of characteristic quantities totally
I_b=I(i-1:i+1,j-1:j+1)-I(i,j)（3）
Extraction Subcellular Localization image characteristic step described in the step 3 comprises,
The abutment points difference counts the pixel number that this each class of subcellular fraction image comprises after calculating,
Secondly, according to the matrix of differences of subcellular fraction image slices vegetarian refreshments, this point is sued for peace respectively greater than the minus difference of zero-sum, ask again the absolute value sum of this difference, as shown in Equation 4,
Then add up respectively each class subcellular fraction image slices vegetarian refreshments greater than zero value summation and minus value summation, can obtain the variance of each class again according to the absolute value sum of each pixel difference in each class at last, so altogether obtain 36 stack features amounts,
sum1=∑(I_b＞0)
sum2=∑(I_b<0)（4）
sum3=sum1-sum2
Like this by after cumulative, sum1 and the sum2 of subcellular fraction image slices vegetarian refreshments in every kind of situation, and the variance of sum3 add total number of subcellular fraction image slices vegetarian refreshments in every kind of situation, can extract altogether the 9x4=36 stack features amount of subcellular fraction image.
The step for the training prediction of Subcellular Localization image described in the step 4 comprises,
Adopt fluorescence antibody or other to visit endogenous protein or the key element of the specific cells device that comprises 10 class organelles of agent detection, every width of cloth image is had an additional specific stain 4' of usefulness DNA, the cell image that 6-diamidino-2-phenylindone (DAPI) is redyed, the nucleus position of each cell in this stain saliency maps picture
The 502 secondary endogenous positioning images of collecting are fixing HeLa cell image, take from 60 multiplying powers under the oil immersion, and this image is 8 gray level images of 768x512 pixel, and every width of cloth contains nearly 13 cells.
The present invention is by extracting feature to each secondary fluorescence microscope images, utilize support vector machine (SVM) algorithm the subcellular fraction image is classified and to predict, the results show, it is very fast that the present invention extracts Subcellular Localization characteristics of image algorithm arithmetic speed, recognition correct rate is high, and also higher to the accuracy rate of subcellular fraction Images Classification.
Description of drawings
The subcellular fraction image of passing threshold differentiation in Fig. 1 embodiment of the invention.
The sample image of the protein of 10 class endogenous expressions among Fig. 2 the present invention.
Fig. 3 feature extraction algorithm of the present invention is to endogenous (Endogenously) Subcellular Localization image set classification accuracy rate table.
Original Subcellular Localization image in Fig. 4 embodiment of the invention.
Pretreated Subcellular Localization image in Fig. 5 embodiment of the invention.
Pixel is adjacent a difference statistical model among Fig. 6 the present invention.
Embodiment
The present invention proposes a kind of algorithm of subcellular fraction image characteristics extraction, each secondary subcellular fraction image is carried out feature extraction, the subcellular fraction image is classified and locate according to the feature of extracting.This algorithm does not need image is carried out cutting, not only can process fast all subcellular fraction images, and has very high accuracy.The present invention realizes can be by the following technical programs:
1, at first the subcellular fraction image is carried out pre-service.To each secondary subcellular fraction image, calculate its pixel value greater than the average u of all pixels of 30, then value being composed less than the pixel value of the pixel of u-30 is zero, obtains pretreated subcellular fraction image.
2, calculating subcellular fraction image slices vegetarian refreshments is adjacent difference a little.In step 1, obtain in the pretreated image, to each pixel value greater than zero pixel, the pixel value difference of 8 pixels that calculating is adjacent, we are labeled as 1 to the abutment points greater than this pixel point value, otherwise be labeled as 0, the number that then is labeled as 1 abutment points around this pixel incorporates this pixel as 9 classes into.
3, extract feature.By step 2, we can count the pixel number that this each class of subcellular fraction image comprises.Secondly, we are according to the matrix of differences of subcellular fraction image slices vegetarian refreshments, this point is sued for peace respectively greater than the minus difference of zero-sum, ask again the absolute value sum of this difference, then we add up each class subcellular fraction image slices vegetarian refreshments greater than zero value summation and minus value summation, again according to the absolute value sum of each pixel difference in each class, can obtain the variance of each class at last, we obtain 36 stack features amounts altogether like this.
4, conclusion.36 groups of subcellular fraction image feature amount extracting according to the present invention are utilized algorithm of support vector machine, and image is trained prediction, the results show, and it is very fast that the present invention extracts subcellular fraction characteristics of image algorithm arithmetic speed, and accuracy is higher simultaneously.
The preprocess method that the present invention adopts:
At first, be illustrated in figure 4 as original Subcellular Localization image, filter out subcellular fraction image intermediate value greater than 30 pixel value, and calculate its average u, shown in the following formula 1:
Sum?Value=∑(I>30).（1）
u=mean(Sum?Value).
Then this sub-picture intermediate value being composed less than the pixel value of u-30 is zero, shown in the following formula 2, the image that obtains as shown in Figure 5:
I=0,if(I<u-30)（2）
The calculating of the present invention design is in abutting connection with the algorithm of point value:
According to the image I that obtains after the pre-service, its pixel value scope is 0, u-30～255.In image I〉0 pixel, calculate the difference of itself and abutment points, shown in the following formula 3.Difference greater than zero be labeled as 1, other be labeled as 0.We can obtain following 9 kinds of model matrixs, are illustrated in fig. 6 shown below.A kind of situation in each pixel corresponding diagram 6 obtains (0)-(8) 9 kinds of characteristic quantities totally.
I_b=I(i-1:i+1,j-1:j+1)-I(i,j)（3）
The present invention designs and extracts the characteristics of image algorithm:
After the abutment points difference was calculated, we can count the pixel number that this each class of subcellular fraction image comprises.Secondly, we sue for peace respectively greater than the minus difference of zero-sum to this point according to the matrix of differences of subcellular fraction image slices vegetarian refreshments, ask the absolute value sum of this difference again, shown in the following formula 4.Then we add up respectively each class subcellular fraction image slices vegetarian refreshments greater than zero value summation and minus value summation, last again according to the absolute value sum of each pixel difference in each class, can obtain the variance of each class, we obtain 36 stack features amounts altogether like this.
sum1=∑(I_b>0)
sum2=∑(I_b<0)（4）
sum3=sum1-sum2
Like this by after cumulative, we can draw among Fig. 4 sum1 and the sum2 of subcellular fraction image slices vegetarian refreshments in every kind of situation, and the variance of sum3, adding total number of subcellular fraction image slices vegetarian refreshments in every kind of situation, we can extract altogether the 9x4=36 stack features amount of subcellular fraction image like this.
The algorithm of image training prediction of the present invention:
The used image set of the present invention is to visit endogenous protein or the key element (10 class organelle) of the specific cells device of agent detection with fluorescence antibody or other.Every width of cloth image is had an additional specific stain 4' of usefulness DNA, the cell image that 6-diamidino-2-phenylindone (DAPI) is redyed, the nucleus position of each cell in this stain energy saliency maps picture, we have collected 502 secondary endogenous positioning images at last.All all is the HeLa cell image of fixing, and takes from 60 multiplying powers under the oil immersion.These all are 8 gray level images of 768x512 pixel, and every width of cloth contains nearly 13 cells.This data set can download from the LOCATE site and obtain (
Http:// locate.imb.uq.edu.au/).
The time test of this experiment is to be that the computer of Intel (R) Xeon (R) 2.00GHz is finished at CPU, and the version of Matlab is MatlabR2010b.
Process is to after the pre-service of subcellular fraction image, utilize extraction characteristics algorithm that the present invention proposes to 503 just feature extractions of secondary subcellular fraction image, 6 minutes 12 seconds this time spent in step of feature extraction, then 36 stack features that extract are carried out normalized after, just can train and predict classification to it.That use here is algorithm of support vector machine (SVM), from 503 secondary subcellular fraction images, extract immediately 400 sub-pictures as svm classifier device training sample at every turn, other 103 sub-pictures are as the forecast sample of sorter, circulate altogether 1000 times, the Average Accuracy that obtains at last classifying is 96.8%.That wherein the realization of SVM is adopted is libsvm tool box (libsvm-mat). this software package can freely obtain at http://www.csie.ntu.edu.tw/ ~ cjlin.
The present invention, we have proposed threshold value in abutting connection with statistical method, its essence is image is carried out that thresholding is processed and calculates pixel quantity above threshold value by a given threshold value.As shown in Figure 1, endoplasmic reticulum (a) and microtubule skeleton (b) image passing threshold (a ' and b ') are so that its intensity is shown as white greater than the pixel of u-30, and u is the mean value of every width of cloth cell image pixel intensity.Although the quality of image a and image b is similar with vision, then more can make a distinction from image a ' and b '.Image a ' has comprised more entity white portion, and image b ' has then shown more internal waviness and emergence edge.
In order to test the effect of this algorithm, we have created a subcellular fraction fluorescence microscope images collection, and each Subcellular Localization image set approximately comprises 50 sub-pictures.This image set is to visit endogenous protein or key element (the 10 class organelles of the specific cells device of agent detection with fluorescence antibody or other, 503 sub-pictures), sample image (engineer's scale is 10 microns) as shown in Figure 2, it represents respectively (a) microtubule (Microtuble), (b) golgiosome (Golgi), (c) plasma membrane (Plasma membrane), (d) actin cytoskeleton (Actin cytoskeleton), (e) nuclear (Nucleus), (f) endosome (Endosome), (g) endoplasmic reticulum (ER), (h) mitochondria (Mitochondria), (i) peroxisome (Peroxisome), (j) lysosome (Lysosome).Method of the present invention shows, need not in the situation of image cropping, under Matlab (R2010b) environment, utilize algorithm of support vector machine, in the situation of cycle calculations 1000 times, to endogenous (Endogenously) Subcellular Localization image set distinguish and when classifying average accuracy can reach 96.8%, experimental result is better than the method that proposes in [3], detailed classified information as shown in Figure 3, chart Central Asia cytological map image set is 10 classes, can find out classification accuracy rate and the wrong minute rate thereof of each class subcellular fraction image, can find out also that by figure this algorithm is to Endosome in this 10 class subcellular fraction image, ER, Mitochondria, Nucleus, PM, Actin-Cytoskeleton, Microtubule and Peroxisome8 class image set classification accuracy rate are higher, and wrong minute rate is less mutually.
Fig. 4 is the secondary original Subcellular Localization image in the Endogenously image set, and its pixel value scope is 0-255.Through first step pre-service, at first calculate all pixel values greater than 30 pixel average u, be pixel value in the whole image 0 less than the tax of u-30 then, not other do not change, obtaining Fig. 5 is pretreated image.Then pixel among Fig. 5, if pixel value is greater than zero, then calculate its adjacent pixels matrix of differences, difference is labeled as 1 greater than zero, otherwise then be labeled as 0, as shown in Figure 6, it is altogether 9 classes that all pixels can be classified as (0)-(8), then can count the number of pixels of this each class of sub-picture.According to each pixel in abutting connection with matrix of differences, count respectively it greater than zero value sum sum1 and minus value sum sum2, and absolute value sum sum3, as shown in Equation 4 shown in.Can be obtained again the variance of each stack features by sum3.And then we obtain 36 stack features amounts.The characteristic quantity that extracts according to our algorithm at last utilizes algorithm of support vector machine, and the Subcellular Localization image is trained prediction, this process circulation 1000 times.Experimental result shows that feature extraction algorithm arithmetic speed of the present invention is very fast, and the Average Accuracy of the Subcellular Localization Images Classification that creates among the present invention is reached 96.8%.
Claims (5)
1. the Subcellular Localization method based on fluorescence microscope images is characterized in that, may further comprise the steps:
Step 1, at first the Subcellular Localization image based on fluorescence microscope images is carried out pre-service, to each secondary Subcellular Localization image, calculate its pixel value greater than the average u of all pixels of 30, then value being composed less than the pixel value of the pixel of u-30 is zero, obtain pretreated Subcellular Localization image
Step 2, calculating subcellular fraction framing pixel is adjacent difference a little, namely in step 1, obtain in the pretreated Subcellular Localization image, to each pixel value greater than zero pixel, the pixel value difference of 8 pixels that calculating is adjacent, abutment points greater than this pixel point value is labeled as 1, otherwise is labeled as 0, the number that then is labeled as 1 abutment points around this pixel incorporates this pixel as 9 classes into;
Step 3, extract the Subcellular Localization characteristics of image, namely by step 2, at first count the pixel number that this each class of Subcellular Localization image comprises, secondly, according to the matrix of differences model of Subcellular Localization image slices vegetarian refreshments, add up each class Subcellular Localization image slices vegetarian refreshments greater than zero value summation and minus value summation, again according to the absolute value sum of each pixel difference in each class, can obtain the variance of each class at last, so altogether obtain 36 stack features amounts;
Step 4 according to 36 groups of Subcellular Localization image feature amount extracting, is utilized algorithm of support vector machine, and the Subcellular Localization image is trained prediction, realizes the Subcellular Localization based on fluorescence microscope images.
2. the Subcellular Localization method based on fluorescence microscope images as claimed in claim 1 is characterized in that, the Subcellular Localization image carried out pre-treatment step comprise described in the step 1,
At first, for the Subcellular Localization image, filter out subcellular fraction image intermediate value greater than 30 pixel value, and calculate its average u, as shown in Equation 1:
Sum?Value=∑(I>30).(1)
u=mean(Sum?Value).
Then this image intermediate value being composed less than the pixel value of u-30 is zero, as shown in Equation 2,
I=0,if(I<u-30)（2）。
3. the Subcellular Localization method based on fluorescence microscope images as claimed in claim 1 is characterized in that, the step that the subcellular fraction framing pixel described in the step 2 is adjacent difference a little comprises,
According to the Subcellular Localization image that obtains after the pre-service, its pixel value scope is 0, u-30～255, in image I〉0 pixel, calculate the difference of itself and abutment points, as shown in Equation 3, difference greater than zero be labeled as 1, other be labeled as 0, obtain 9 kinds of model matrixs in the following table, a kind of situation in the corresponding table of each pixel obtains (0)-(8) 9 kinds of characteristic quantities totally
I_b=I(i-1:i+1,j-1:j+1)-I(i,j)（3）
4. the Subcellular Localization method based on fluorescence microscope images as claimed in claim 1 is characterized in that, the extraction Subcellular Localization image characteristic step described in the step 3 comprises,
The abutment points difference counts the pixel number that this each class of subcellular fraction image comprises after calculating,
Secondly, according to the matrix of differences of subcellular fraction image slices vegetarian refreshments, this point is sued for peace respectively greater than the minus difference of zero-sum, ask again the absolute value sum of this difference, as shown in Equation 4,
Then add up respectively each class subcellular fraction image slices vegetarian refreshments greater than zero value summation and minus value summation, can obtain the variance of each class again according to the absolute value sum of each pixel difference in each class at last, so altogether obtain 36 stack features amounts,
sum1=∑(I_b>0)
sum2=∑(I_b<0)（4）
sum3=sum1-sum2
Like this by after cumulative, sum1 and the sum2 of subcellular fraction image slices vegetarian refreshments in every kind of situation, and the variance of sum3 add total number of subcellular fraction image slices vegetarian refreshments in every kind of situation, can extract altogether the 9x4=36 stack features amount of subcellular fraction image.
5. the Subcellular Localization method based on fluorescence microscope images as claimed in claim 1 is characterized in that, the step for the training prediction of Subcellular Localization image described in the step 4 comprises,
Adopt fluorescence antibody or other to visit endogenous protein or the key element of the specific cells device that comprises 10 class organelles of agent detection, every width of cloth image is had an additional specific stain 4' of usefulness DNA, the cell image that 6-diamidino-2-phenylindone (DAPI) is redyed, the nucleus position of each cell in this stain saliency maps picture
The 502 secondary endogenous positioning images of collecting are fixing HeLa cell image, take from 60 multiplying powers under the oil immersion, and this image is 8 gray level images of 768x512 pixel, and every width of cloth contains 13 cells.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201210588110.5A CN103020488B (en) | 2012-12-29 | 2012-12-29 | A kind of Subcellular Localization method based on fluorescence microscope images |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201210588110.5A CN103020488B (en) | 2012-12-29 | 2012-12-29 | A kind of Subcellular Localization method based on fluorescence microscope images |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN103020488A true CN103020488A (en) | 2013-04-03 |
| CN103020488B CN103020488B (en) | 2016-10-05 |
Family
ID=47969086
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201210588110.5A Expired - Fee Related CN103020488B (en) | 2012-12-29 | 2012-12-29 | A kind of Subcellular Localization method based on fluorescence microscope images |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103020488B (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104881857A (en) * | 2014-02-28 | 2015-09-02 | 中国长江三峡集团公司中华鲟研究所 | Image analysis calculating method for fish oocyte polar nucleus deviant |
| CN107513552A (en) * | 2017-10-21 | 2017-12-26 | 云南中烟工业有限责任公司 | The detection method that a kind of cigarette smoke influences on the cell expression quantity of aquaporin 5 |
| CN107607511A (en) * | 2017-10-21 | 2018-01-19 | 云南中烟工业有限责任公司 | The detection method that a kind of cigarette smoke influences on the Subcellular Localization of aquaporin 5 |
| CN108550148A (en) * | 2018-04-13 | 2018-09-18 | 重庆大学 | Nucleus in histotomy micro-image divides automatically and classifying identification method |
| CN112184696A (en) * | 2020-10-14 | 2021-01-05 | 中国科学院近代物理研究所 | Method and system for counting cell nucleus and cell organelle and calculating area of cell nucleus and cell organelle |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101176116A (en) * | 2005-05-13 | 2008-05-07 | 三路影像公司 | Methods of chromogen separation-based image analysis |
| US7430319B2 (en) * | 2001-12-19 | 2008-09-30 | Fuji Xerox Co., Ltd. | Image collating apparatus for comparing/collating images before/after predetermined processing, image forming apparatus, image collating method, and image collating program product |
-
2012
- 2012-12-29 CN CN201210588110.5A patent/CN103020488B/en not_active Expired - Fee Related
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7430319B2 (en) * | 2001-12-19 | 2008-09-30 | Fuji Xerox Co., Ltd. | Image collating apparatus for comparing/collating images before/after predetermined processing, image forming apparatus, image collating method, and image collating program product |
| CN101176116A (en) * | 2005-05-13 | 2008-05-07 | 三路影像公司 | Methods of chromogen separation-based image analysis |
Non-Patent Citations (3)
| Title |
|---|
| LORIS NANNI等: "Fusion of systems for automated cell phenotype image classification", 《EXPERT SYSTEMS WITH APPLICATIONS》, vol. 37, 31 December 2010 (2010-12-31) * |
| NICHOLAS A HAMILTON等: "Fast automated cell phenotype image classification", 《BMC BIOINFORMATICS》, 30 March 2007 (2007-03-30) * |
| 张树波等: "蛋白质亚细胞定位预测的机器学习方法", 《计算机科学》, vol. 36, no. 4, 15 April 2009 (2009-04-15) * |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104881857A (en) * | 2014-02-28 | 2015-09-02 | 中国长江三峡集团公司中华鲟研究所 | Image analysis calculating method for fish oocyte polar nucleus deviant |
| CN104881857B (en) * | 2014-02-28 | 2017-06-06 | 中国长江三峡集团公司中华鲟研究所 | A kind of image analysis calculation method of fish-egg mother cell polar core deviant |
| CN107513552A (en) * | 2017-10-21 | 2017-12-26 | 云南中烟工业有限责任公司 | The detection method that a kind of cigarette smoke influences on the cell expression quantity of aquaporin 5 |
| CN107607511A (en) * | 2017-10-21 | 2018-01-19 | 云南中烟工业有限责任公司 | The detection method that a kind of cigarette smoke influences on the Subcellular Localization of aquaporin 5 |
| CN107513552B (en) * | 2017-10-21 | 2021-02-02 | 云南中烟工业有限责任公司 | Method for detecting influence of cigarette smoke on aquaporin 5 cell expression quantity |
| CN108550148A (en) * | 2018-04-13 | 2018-09-18 | 重庆大学 | Nucleus in histotomy micro-image divides automatically and classifying identification method |
| CN112184696A (en) * | 2020-10-14 | 2021-01-05 | 中国科学院近代物理研究所 | Method and system for counting cell nucleus and cell organelle and calculating area of cell nucleus and cell organelle |
| CN112184696B (en) * | 2020-10-14 | 2023-12-29 | 中国科学院近代物理研究所 | Cell nucleus and organelle counting and area calculating method and system thereof |
Also Published As
| Publication number | Publication date |
|---|---|
| CN103020488B (en) | 2016-10-05 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109034159B (en) | Image information extraction method and device | |
| CN105512683B (en) | Object localization method and device based on convolutional neural networks | |
| Phansalkar et al. | Adaptive local thresholding for detection of nuclei in diversity stained cytology images | |
| Ersoy et al. | HEp-2 cell classification in IIF images using shareboost | |
| CN103020488B (en) | A kind of Subcellular Localization method based on fluorescence microscope images | |
| CN102663431B (en) | Image matching calculation method on basis of region weighting | |
| CN107610177B (en) | The method and apparatus of characteristic point is determined in a kind of synchronous superposition | |
| JP2012104111A (en) | Method and system for detecting multi-view human face | |
| CN104036259B (en) | Human face similarity degree recognition methods and system | |
| CN111461039A (en) | Landmark identification method based on multi-scale feature fusion | |
| Yin et al. | Accurate and robust text detection: A step-in for text retrieval in natural scene images | |
| Li et al. | Mix-up augmentation for oracle character recognition with imbalanced data distribution | |
| Newberg et al. | Automated analysis of human protein atlas immunofluorescence images | |
| Dornaika et al. | Object-centric contour-aware data augmentation using superpixels of varying granularity | |
| Sainin et al. | Malaysian medicinal plant leaf shape identification and classification | |
| CN113822134A (en) | Instance tracking method, device, equipment and storage medium based on video | |
| Cheng et al. | Discriminative segmentation of microscopic cellular images | |
| Noman et al. | Improving accuracy and efficiency in seagrass detection using state-of-the-art AI techniques | |
| Huh et al. | Efficient framework for automated classification of subcellular patterns in budding yeast | |
| Piriyothinkul et al. | Detecting text in manga using stroke width transform | |
| Satish et al. | Edge assisted fast binarization scheme for improved vehicle license plate recognition | |
| Hu et al. | Application research of automatic garbage sorting based on TensorFlow and OpenCV | |
| CN112802026A (en) | Deep learning-based real-time traffic scene semantic segmentation method | |
| Vieira et al. | Human epithelial type 2 (HEp-2) cell classification by using a multiresolution texture descriptor | |
| Tatke et al. | Hybrid approach of garbage classification using computer vision and deep learning |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20161005 Termination date: 20201229 |
|
| CF01 | Termination of patent right due to non-payment of annual fee |