CN103186791B - Particle scatter diagram classifying method based on iteration - Google Patents
Particle scatter diagram classifying method based on iteration Download PDFInfo
- Publication number
- CN103186791B CN103186791B CN201110459264.XA CN201110459264A CN103186791B CN 103186791 B CN103186791 B CN 103186791B CN 201110459264 A CN201110459264 A CN 201110459264A CN 103186791 B CN103186791 B CN 103186791B
- Authority
- CN
- China
- Prior art keywords
- particle
- iteration
- inactive area
- statistical information
- scatter diagram
- 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.)
- Expired - Fee Related
Links
- 239000002245 particle Substances 0.000 title claims abstract description 135
- 238000010586 diagram Methods 0.000 title claims abstract description 46
- 238000000034 method Methods 0.000 title claims abstract description 24
- 210000000265 leukocyte Anatomy 0.000 description 12
- 230000005484 gravity Effects 0.000 description 8
- 210000004027 cell Anatomy 0.000 description 4
- 238000001514 detection method Methods 0.000 description 4
- 210000004369 blood Anatomy 0.000 description 3
- 239000008280 blood Substances 0.000 description 3
- 230000004323 axial length Effects 0.000 description 2
- 238000006243 chemical reaction Methods 0.000 description 2
- 230000002902 bimodal effect Effects 0.000 description 1
- 210000000601 blood cell Anatomy 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 239000012141 concentrate Substances 0.000 description 1
- 230000007812 deficiency Effects 0.000 description 1
- 238000003064 k means clustering Methods 0.000 description 1
- 239000007788 liquid Substances 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000005693 optoelectronics Effects 0.000 description 1
- 230000007115 recruitment Effects 0.000 description 1
- 238000000926 separation method Methods 0.000 description 1
Landscapes
- Image Analysis (AREA)
Abstract
The invention discloses a particle scatter diagram classifying method based on iteration. The method comprises the following steps: A, particles in a particle scatter diagram are classified so as to acquire statistical information of classification of each particle; B, an invalid area of the particles is estimated through the statistical information of the classification of the each particle; C, particles in the invalid area are separated so as to acquire statistical information of classification of each particle; and D, whether the difference of the statistical information of the classification of the each particle after the iteration and the statistical information of the classification of the each particle after the former iteration is smaller than a threshold value or not is adjusted: if the difference of the statistical information of the classification of the each particle after the iteration and the statistical information of the classification of the each particle after the former iteration is smaller than the threshold value, iteration is stopped and classification results are output; and otherwise, the step A is turned to and iteration classification processing is started again. The particle scatter diagram classifying method based on the iteration reduces difficulties, brought by modeling through people, of accurately estimating the invalid area, effectively improves stability of particle scatter diagram classifying and accuracy of the particle scatter diagram classifying, and has the advantages of being easy to realize.
Description
Technical field
The present invention relates to detection of particles technical field, especially relate to a kind of method that particle scatter diagram nicety of grading is improved by iteration.
Background technology
Detection of particles instrument is used for carrying out classified statistic, such as blood cell analyzer to various particles, and its major function is exactly to provide the number of various types of cells in human body, is basic function wherein to the counting of all kinds of leukocyte.Make after agent treated, by the coated leukocyte of liquid one by one pass through detection zone（Flow chamber）, laser is irradiated on single cell, by collecting refraction or the scattering light of different angles（It is usually forward scattering light FSC, side scattered light SSC）Pass through opto-electronic conversion and AD conversion again, obtain one group of 2-D data of corresponding cell, this 2-D data is mapped in coordinate system, position in two-dimensional coordinate system for this cell can be obtained, all leukocyte in sample are mapped to can obtain the leukocyte scattergram of two dimension, abbreviation scatterplot on coordinate.
In scatterplot, together, inhomogeneous leukocyte is disconnected from each other, as shown in Figure 1 for of a sort leukocyte recruitment.Multiple regions are divided on scatterplot, is classified as same class falling in the leukocyte of uniform areas, and counts the particle data in these classifications and percentage ratio, the one-tenth for analyzing tested sample is grouped into.
Typically require in addition to the particle of the effective needs classification shown in Fig. 1, in the particle scatter diagram of classification, also include some invalid information, i.e. blood shadow part in Null Spot, such as cellanalyzer leukocyte scatterplot, inactive area as shown in Figure 2.These Null Spots can produce impact to the result of particle classifying, especially when Null Spot is more, the precision of classification can be produced and significantly affect, even result in the wrong identification of whole particle group.Often there is dispersibility and unstability in these Null Spots, that is, for different inputs, the position of inactive area, shapes and sizes are all slightly different, and is therefore only capable of determining invalid particle area by the method estimated.
The sorting technique of traditional particle scatter diagram typically first estimates inactive area, is separated, then carries out the classification of particle group.In view of the dispersibility of Null Spot distribution mentioned before and unstability, direct estimation Null Spot distributed areas are more difficult, and estimated result accuracy is relatively low, thus leading to mistake to be classified.
Content of the invention
The present invention is exactly to make up the deficiency of above-mentioned sorting technique it is proposed that a kind of method improving particle scatter diagram nicety of grading by iteration, using the teaching of the invention it is possible to provide stable, accurate classification results.
The present invention adopts the following technical scheme that realization：A kind of particle scatter diagram sorting technique based on iteration, it includes step：
A, particle in particle scatter diagram is classified, obtain the statistical information of each particle classifying；
B, using each particle classifying statistical information estimate particle inactive area；
C, by the separate particles in inactive area, and obtain the statistical information of each particle classifying；
D, judge the statistical information of each particle classifying and last iteration after current iteration after the difference of statistical information of each particle classifying whether be less than threshold value, if so, then terminate iteration output category result, otherwise proceed to step A and start Iterative classification again and process.
Wherein, before step A, methods described also includes step：According to a preliminary estimate in particle scatter diagram particle inactive area, and by the separate particles in inactive area.
Wherein, described step B is to set up model according to the feature of scatterplot, determines the statistical information of each particle classifying and the relation of inactive area in particle scatter diagram, utilizes the inactive area of the statistical information estimation particle of each particle classifying with this.
Wherein, the described step by the separate particles in inactive area is by the population zero setting in inactive area.
In addition, invention additionally discloses a kind of particle scatter diagram categorizing system based on iteration, its described system includes：
Preliminary classification processing module, for the inactive area of particle in particle scatter diagram according to a preliminary estimate, and by the separate particles in inactive area；
Classification information statistical module, for classifying, obtains the statistical information of each particle classifying to particle in particle scatter diagram；
Precise classification processing module, for estimating the inactive area of particle using the statistical information of each particle classifying, by the separate particles in inactive area；
Comparison module, whether the difference for judging the statistical information of each particle classifying after the statistical information of each particle classifying and last iteration after current iteration is less than threshold value, if, then terminate iteration and by classification results output module output category result, otherwise Iterative classification is processed the startup of control tactics Information Statistics module again；
Classification results output module, for output category result after terminating iteration.
Compared with prior art, the present invention has the advantages that：
The invention provides improving, in order to automatic, the method that particle scatter diagram inactive area estimates accuracy and particle classifying precision, reduce artificial modeling with the accurate difficulty estimating inactive area, it is effectively improved stability and the accuracy of particle scatter diagram classification, have and realize simple advantage.
Brief description
Fig. 1 is the schematic diagram of particle group in scatterplot；
Fig. 2 is the schematic diagram of original scatterplot；
Fig. 3 is the schematic flow sheet of the present invention；
Fig. 4 is the schematic diagram of inactive area initial option in scatterplot；
Fig. 5 is the classification schematic diagram accurately estimating scatterplot inactive area；
Fig. 6 is the system structure diagram of the present invention.
Specific embodiment
Null Spot distributed areas are identified as the invalid distribution being often as this kind of particle and not concentrate, and are not enough to classify, or the classification results of this kind of particle no essential meaning when analyzing this particle scatter diagram.In fact, Null Spot is also to be formed by particle, exists between the distribution of Null Spot and other classification and necessarily contact.Therefore, the classification results of effective coverage contribute to defining of inactive area.
As shown in figure 3, specifically, the present invention provides one kind to be applied to particle analyzer, the method in order to improve particle scatter diagram nicety of grading, comprises the steps：
Step S11, according to particle the scattered light signal in different angle of scatterings（As forward scattering light FSC and side scattered light SSC）Generate predecessor scatterplot（Non-classified scatterplot）, than as shown in Figure 2 with the leukocyte scatterplot of cellanalyzer.
The position of particle inactive area in step S12, according to a preliminary estimate particle scatter diagram, and by the separate particles in inactive area.Separation method is such as：By the population zero setting in inactive area.
This step does not need to accurately calculate the position of inactive area it is only necessary to provide approximate region, to reduce the interference to classification results for the inactive area.Taking the leukocyte scatterplot of cellanalyzer as a example：Main impact classification results are blood shadow regions, see inactive area I in Fig. 3.Fig. 3 is that comparatively ideal leukocyte particle scatter diagram classifies schematic diagram it can be seen that the particle of inactive area is only intensive in lower left quarter, and other region populations are less, classification results affected less.
Therefore, in step s 12, the estimation of inactive area is only needed generally to exclude the dense particles of lower left quarter.In one embodiment, after leukocyte scatterplot being carried out according to a preliminary estimate, substantially mark off inactive area.Such as, inactive area can tentatively delimit inactive area simply by horizontal linear LA and this two lines of vertical line LB, as shown in Figure 4.The method for determining position of horizontal linear and vertical line can pass through the height of scatterplot respectively and width calculation obtains, but is not limited to the method.
Step S13, on the scatterplot eliminating inactive area, particle is classified（Such as pass through K-Means clustering algorithm）, and obtain the statistical information of each classification., the statistical information of each classification mainly has other statistical information in the position of centre of gravity of each particle classifying and all directions, the such as axial length of cluster group taking cellanalyzer as a example.
Step S14, using each particle classifying calculating statistical information, the estimation position of inactive area of particle, shapes and sizes again, so that it is determined that inactive area.Different to the simple estimation of inactive area from step S12, this estimates the statistical information based on particle classifying, it is possible to obtain more accurately estimate.The statistical information of each particle classifying of scatterplot and the relation of inactive area, need to set up model according to the specific features of scatterplot.Such as, as Fig. 5 shows, the circular arc portion L11 of the boundary line of inactive area can be calculated by the statistical information classified in Fig. 4 lower left region and obtain（Scatterplot lower left quarter is projected to y-axis, form projection histogram, lower-left cluster group and blood shadow compact district below will be formed on the histogram bimodal, find the demarcation line that valley can obtain between this cluster group and dead space, position of centre of gravity and vertical axial length in conjunction with this cluster group can obtain the center of circle and the radius of this circular arc）；The height of gable L12 of the boundary line of inactive area is related to the lower limit of classification above delta-shaped region, and the lower limit of this classification can be calculated by the statistical information of this classification and obtain（Be mutually perpendicular to by two that this cluster is rolled into a ball and cross center of gravity axle length, in conjunction with this classification position of centre of gravity, it is possible to obtain this cluster group lower limit, that is, in y-axis minimum be likely to occur unknown）；The left hypotenuse original position of gable L12 and slope are all related to the statistical information of two classification in the left side（The slope linear correlation of the two classification center of gravity lines in the slope of left hypotenuse and the left side, original position can be calculated by the slope of two classification center of gravity lines of position of centre of gravity and the left side of the two classification any of which classification in the left side and obtain）；The original position of right hypotenuse of gable L12 and slope are all related to the statistical information of surface classification（The position of centre of gravity linear correlation of classification above the original position of right hypotenuse and trigonum, the slope linear correlation of the slope of right hypotenuse and top classification major axis）.Above-mentioned model is only applicable to the scatterplot that the scatterplot shown in Fig. 5 and same particle detection systems generate, and the statistical information of each particle classifying of scatterplot and the relational model of dead space are not limited to the method.
Step S15, according to step D gained information, again will isolate, in original scatterplot, the inactive area that invalid particle is located, reacquire the statistical information of each particle classifying, concrete grammar is with described in step S13.The raising of the accuracy due to estimating to inactive area, the accuracy of classification results is also improved.
Step S16, judge the statistical information of each particle classifying and last iteration after current iteration after the difference of statistical information of each particle classifying whether be less than threshold value, if so, then proceed to step S17, otherwise proceed to step S14.
Step S14 and step S15 can repeat, pass through to improve classification results accuracy, improve the accuracy that inactive area is estimated, pass through again to improve the accuracy that inactive area is estimated, improve the accuracy of classification results, gradually improve nicety of grading in iteration, after the population of each classification and last iteration after this iteration, the difference of the population of each classification is less than a threshold value（Threshold value is a fixed value set in advance）Then it is assumed that classification is stable, iteration output category result are terminated by step S17.
Step S17, termination iteration output category result.
As shown in fig. 6, corresponding, invention additionally discloses a kind of particle scatter diagram categorizing system based on iteration, it includes：Preliminary classification processing module 11 for execution step S12；Classification information statistical module 12 for execution step S13；Precise classification processing module 13 for execution step S14 and step S15, this precise classification processing module 13 carries out after invalid separate particles to the inactive area of accurate estimation, and still again particle is classified by classification information statistical module 12 and obtained with the statistical information of each classification；Comparison module 14 for execution step S16, after the population of its each classification after judging current iteration and last iteration, the difference of the population of each classification is less than threshold value, then think that classification is stable, iteration output category result are terminated by classification results output module 15, otherwise control tactics Information Statistics module 12 is iterated to particle scatter diagram operating again.
The invention provides improving, in order to automatic, the method that particle scatter diagram inactive area estimates accuracy and particle classifying precision, reducing artificial modeling with the accurate difficulty estimating inactive area, being effectively improved stability and the accuracy of particle scatter diagram classification.
The foregoing is only presently preferred embodiments of the present invention, not in order to limit the present invention, all any modification, equivalent and improvement made within the spirit and principles in the present invention etc., should be included within the scope of the present invention.
Claims (6)
1. a kind of particle scatter diagram sorting technique based on iteration is it is characterised in that methods described includes step：
According to a preliminary estimate in particle scatter diagram particle inactive area, and by the separate particles in inactive area；
A, particle in particle scatter diagram is classified, obtain the statistical information of each particle classifying；
B, using each particle classifying statistical information estimate particle inactive area；
C, by the separate particles in inactive area, and obtain the statistical information of each particle classifying；
D, judge the statistical information of each particle classifying and last iteration after current iteration after the difference of statistical information of each particle classifying whether be less than threshold value, if so, then terminate iteration output category result, otherwise proceed to step A and start Iterative classification again and process.
2. the particle scatter diagram sorting technique based on iteration according to claim 1, it is characterized in that, described step B is to set up model according to the feature of scatterplot, determine the statistical information of each particle classifying and the relation of inactive area in particle scatter diagram, the inactive area of the statistical information estimation particle of each particle classifying is utilized with this.
3. the particle scatter diagram sorting technique based on iteration according to claim 1 or claim 2 is it is characterised in that the described step by the separate particles in inactive area is by the population zero setting in inactive area.
4. a kind of particle scatter diagram categorizing system based on iteration is it is characterised in that described system includes：
Preliminary classification processing module, for the inactive area of particle in particle scatter diagram according to a preliminary estimate, and by the separate particles in inactive area, the input of the outfan link sort Information Statistics module of preliminary classification processing module；
Classification information statistical module, for classifying, obtains the statistical information of each particle classifying to particle in particle scatter diagram；
Precise classification processing module, for estimating the inactive area of particle using the statistical information of each particle classifying, by the separate particles in inactive area；
Comparison module, whether the difference for judging the statistical information of each particle classifying after the statistical information of each particle classifying and last iteration after current iteration is less than threshold value, if, then terminate iteration and by classification results output module output category result, otherwise Iterative classification is processed the startup of control tactics Information Statistics module again；
Classification results output module, for output category result after terminating iteration.
5. the particle scatter diagram categorizing system based on iteration according to claim 4, it is characterized in that, described precise classification processing module is to set up model according to the feature of scatterplot, determine the statistical information of each particle classifying and the relation of inactive area in particle scatter diagram, the inactive area of the statistical information estimation particle of each particle classifying is utilized with this.
6. according to claim 4 or 5 the particle scatter diagram categorizing system based on iteration it is characterised in that described is by the population zero setting in inactive area by the separate particles in inactive area.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201110459264.XA CN103186791B (en) | 2011-12-31 | 2011-12-31 | Particle scatter diagram classifying method based on iteration |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201110459264.XA CN103186791B (en) | 2011-12-31 | 2011-12-31 | Particle scatter diagram classifying method based on iteration |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN103186791A CN103186791A (en) | 2013-07-03 |
| CN103186791B true CN103186791B (en) | 2017-02-15 |
Family
ID=48677951
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201110459264.XA Expired - Fee Related CN103186791B (en) | 2011-12-31 | 2011-12-31 | Particle scatter diagram classifying method based on iteration |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103186791B (en) |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102305758A (en) * | 2011-05-19 | 2012-01-04 | 长春迪瑞医疗科技股份有限公司 | Method for quickly and automatically classifying particles and implementation device thereof |
| CN102507417A (en) * | 2011-11-29 | 2012-06-20 | 长春迪瑞医疗科技股份有限公司 | Method for automatically classifying particles |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102279146A (en) * | 2011-03-11 | 2011-12-14 | 桂林优利特医疗电子有限公司 | Blood cell five classification method based on laser sheath flow technology |
-
2011
- 2011-12-31 CN CN201110459264.XA patent/CN103186791B/en not_active Expired - Fee Related
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102305758A (en) * | 2011-05-19 | 2012-01-04 | 长春迪瑞医疗科技股份有限公司 | Method for quickly and automatically classifying particles and implementation device thereof |
| CN102507417A (en) * | 2011-11-29 | 2012-06-20 | 长春迪瑞医疗科技股份有限公司 | Method for automatically classifying particles |
Also Published As
| Publication number | Publication date |
|---|---|
| CN103186791A (en) | 2013-07-03 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109952614B (en) | Biological particle classification system and method | |
| CN105940301B (en) | A kind of stream type cell analyzer and its multidimensional data sorting technique, device | |
| US20080172185A1 (en) | Automatic classifying method, device and system for flow cytometry | |
| CN102507417B (en) | Method for automatically classifying particles | |
| CN101672759B (en) | Classified statistic method and device of particles | |
| CN105203446B (en) | Based on probability distribution cell classification statistical method | |
| CN107389536B (en) | Flow cell particle classification counting method based on density-distance center algorithm | |
| CN104200114B (en) | Flow cytometry data rapid analysis method | |
| CN113188981B (en) | Automatic analysis method of multi-factor cytokine | |
| CN113516105B (en) | Lane detection method and device and computer readable storage medium | |
| CN106228554A (en) | Fuzzy coarse central coal dust image partition methods based on many attribute reductions | |
| CN103364324A (en) | Self-adapted classified counting method for blood cell analyzer | |
| WO2018006631A1 (en) | User level automatic segmentation method and system | |
| CN103345575B (en) | A kind of data flow concept drift detection method and system | |
| CN103679214A (en) | Vehicle detection method based on online area estimation and multi-feature decision fusion | |
| CN109580550A (en) | A kind of classification processing method and its device of leucocyte | |
| Rahman et al. | Automatic identification of abnormal blood smear images using color and morphology variation of RBCS and central pallor | |
| Hagwood et al. | Evaluation of segmentation algorithms on cell populations using CDF curves | |
| CN101762448B (en) | Method for classifying particles and device for detecting particles | |
| CN103186791B (en) | Particle scatter diagram classifying method based on iteration | |
| CN104361224A (en) | Confidence classification method and confidence machine | |
| CN103776751B (en) | For the method for classifying particles of scatter diagram in detection of particles instrument | |
| Wei et al. | Automatic counting method for complex overlapping erythrocytes based on seed prediction in microscopic imaging | |
| CN104700422A (en) | Method for automatically segmenting bonded red blood cells and white blood cells in urinary sediment image | |
| CN106501160A (en) | A kind of method for classifying particles and particle classifying device |
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 | ||
| TR01 | Transfer of patent right | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20170710 Address after: 518000 Guangdong city of Shenzhen province Nanshan District Xili Street Xueyuan Road No. 1001 Nanshan Chi Park A4 Building 9 floor Patentee after: SHENZHEN LANDWIND BIOLOGICAL ENGINEERING CO.,LTD. Address before: Futian District King Road Shenzhen city Guangdong province 518000 No. 81 country garden building E 601 Patentee before: SHENZHEN LANDWIND INDUSTRY Co.,Ltd. |
|
| TR01 | Transfer of patent right | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20221223 Address after: Room 101, Building B15, Life Science Industrial Park, No. 140, Jinye Avenue, Sanxi Community, Kuichong Street, Dapeng New District, Shenzhen, Guangdong 518000 Patentee after: Shenzhen Lanyun Biomedical Technology Co.,Ltd. Address before: 518000 9, A4 building, Nanshan Zhiyuan 1001, Xue Yuan Avenue, Xili street, Nanshan District, Shenzhen, Guangdong. Patentee before: SHENZHEN LANDWIND BIOLOGICAL ENGINEERING CO.,LTD. |
|
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20170215 |