CN109754398A - A kind of computer based myeloplast automark method and system - Google Patents
A kind of computer based myeloplast automark method and system Download PDFInfo
- Publication number
- CN109754398A CN109754398A CN201910031626.1A CN201910031626A CN109754398A CN 109754398 A CN109754398 A CN 109754398A CN 201910031626 A CN201910031626 A CN 201910031626A CN 109754398 A CN109754398 A CN 109754398A
- Authority
- CN
- China
- Prior art keywords
- myeloplast
- cell
- image
- classification
- label
- 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
Links
Landscapes
- Investigating Or Analysing Biological Materials (AREA)
Abstract
The invention discloses a kind of computer based myeloplast automark methods, including read cell text information file；Cell image CellImage is inputted into preset disaggregated model；Myeloplast classification figure layer image generates；The corresponding cell number of collect statistics label and all cell number mesh etc., invention additionally discloses a kind of automatic tagging systems of computer based myeloplast.The present invention can complete slide from digitlization to the multiple functions such as Classification and Identification and multi-layer image visualization, both electronics diagosis data source problem had been can solve, the meaning of myeloplast tagging system can also be realized to the greatest extent, pathologist is allowed to complete related pathologies diagnostic work dependent on computer, different types of leucocyte multi-layer image visualization interface that computer mark comes out also can satisfy demand different during pathologist diagnoses, its fast accurate is assisted to complete pathological diagnosis.
Description
Technical field
Fig. 1 is that the present invention is based on the flow charts of the myeloplast automark method of computer；
Fig. 2 is that the present invention is based on the module diagrams of the automatic tagging system of the myeloplast of computer.
Specific embodiment
Following will be combined with the drawings in the embodiments of the present invention, and technical solution in the embodiment of the present invention carries out clear, complete
Site preparation description, it is clear that described embodiments are only a part of the embodiments of the present invention, instead of all the embodiments.It is based on
Embodiment in the present invention, it is obtained by those of ordinary skill in the art without making creative efforts all other
Embodiment shall fall within the protection scope of the present invention.
Referring to FIG. 1-2, specific embodiment 1: the invention provides the following technical scheme: a kind of computer based bone
Marrow leucocyte automark method, includes the following steps:
Step 1: reading cell text information file, cell image CellImag is determined according to (x, y, R)；
Step 2: cell image CellImage is inputted preset disaggregated model, obtains the cell and belong to marrow difference leucocyte
Probability tables, the corresponding label class of maximum probability and its probability value are stored according to (x, y, R, label, Probablity) group
In the text structural information knitted；
Step 3: myeloplast classification figure layer image generates, such is belonged to according to the generation of the label label of default disaggregated model
The other all cells of same class are corresponded at it and use color at position (x, y, R) by other particular color ColorLabel
ColorLabel is marked with circle, generates classification figure layer image LabelLayerImage；
Step 4: the corresponding cell number of collect statistics label and all cell number mesh, generate different classes of myeloplast
Accounting information, numerical value visualization and carry out histogram visualization, by not in the cellular informatics of normal rates range class with red
Color box is marked.
It may include different myeloplast types that preset disaggregated model, which carries out classification to cell, in the step two
Model, prototype network structure and leucocyte type can make corresponding change according to demand.Wherein, preset classification in step 2
The network structure of model uses InceptionV3 to carry out transfer learning, and corresponding tag along sort includes: granulocyte, lymph
Cell, monocyte, red blood cell, eosinophil, basophilic granulocyte, thick liquid cell.In the training set of default disaggregated model
Including granulocyte: 12000, lymphocyte: 10000, monocyte: 2000, red blood cell: 6004, eosinophil: 2014,
Basophilic granulocyte: 300, thick liquid cell: 702.Data enhancement methods include: mirror image, [0,45,90,135,180] rotation, translation.
By data enhancing so that the amount of images of each classification increases, online Enhancement Method: mirror image, 0- is used in the training process
360 ° of rotations, translation, noise prevention training over-fitting.Retain the best model of verifying collection effect as default disaggregated model.
A kind of automatic tagging system of computer based myeloplast, the system include slide digital module, number
Change image slice module, cell image fixation and recognition module, slide cellular informatics Macro or mass analysis module and visualization model:
The slide digital module: the pathology of each actual patient is obtained to scan multiplying power ScanPlan by scanner
The digitized image DigitalImage of slice, digitized image is to scan specific format storage；Meanwhile slide digital module
The middle common multiplying power of ScanPlan is 100 times of object lens.
The digitized image stripping and slicing module: by reading digitized image DigitalImage, by digitized image
The fritter Blocks that DigitalImage is cut into 2048*2048 facilitates the subsequent progress parallel processing of computer to improve efficiency；
The cell image fixation and recognition module: Blocks is inputted into queue to be processed, computer is each time from team to be processed
The automatic ProperNum image data of reading of column is automatically positioned dividing method by the cell of myeloplast and quickly identifies marrow
Leucocyte, and the myeloplast of identification is input in trained corresponding neural network model and obtains its correspondence classification information
And its probability, and relevant information is arrived into CellsPosition message structure according to (x, y, R, label, Probablity) storage
In text；Meanwhile the preset classification mould in cell image fixation and recognition module in the value and claim 2 of ProperNum
Type complexity is related.
The slide cellular informatics Macro or mass analysis module: CellsPosition message structure text is read, according to wherein
The corresponding cell number of (x, y, R, label, Probablity) collect statistics label and all cell number mesh generate different
The accounting information of classification myeloplast；
The visualization model: numerical value visualization is carried out according to slide cellular informatics Macro or mass analysis module results and carries out column
Figure visualization, will not be marked in the cellular informatics of normal rates range class with red boxes；Simultaneously by classification figure layer figure
Picture LabelLayerImage and raw digitized image DigitalImage progress is compound, can check unitary class according to selection
Other myeloplast tag image either multi-class myeloplast tag image.
Wherein, ProperNum value is 10 in cell image fixation and recognition module.Classification figure layer image in visualization model
The sum of LabelLayerImage is 7, granulocyte, lymphocyte, monocyte, red blood cell, the acidophilia of corresponding disaggregated model
Granulocyte, basophilic granulocyte, thick liquid cell.It can select to check which type leucocyte by the button at interface or check simultaneously
Multiple type leucocytes.
Specific embodiment 2:
The network structure of preset disaggregated model uses MobileNet, corresponding tag along sort packet in the step two
It includes: granulocyte, lymphocyte, monocyte, red blood cell.Include granulocyte in the training set of default disaggregated model: 14314, drenching
Bar cell: 10000, monocyte: 2000, red blood cell: 6004.Data enhancement methods include: mirror image, [0,45,90,135,
180] it rotates, translate.Enhanced by data so that the amount of images of each classification increases, reservation verifying collects the best model of effect
As default disaggregated model.
ProperNum value is 20 in the cell image fixation and recognition module.Classification figure layer figure in visualization model
As the sum of LabelLayerImage is 4, granulocyte, lymphocyte, monocyte, the red blood cell of corresponding disaggregated model.
It should be understood that the above-mentioned description for preferred embodiment is more detailed, can not therefore be considered to this
The limitation of invention patent protection range, the present invention can also have other numerous embodiments, and input picture can be any multiplying power figure
Picture, other relevant parameters can be adjusted with figure, and to can be any one network model either traditional for default disaggregated model
Disaggregated model, wherein the picture number of parallel processing is according to the complicated dynamic behaviour of disaggregated model.Those skilled in the art are in this hair
Under bright enlightenment, in the case where not departing from the ambit that the claims in the present invention are protected, replacement or deformation can also be made, is fallen
Enter within protection scope of the present invention, it is of the invention range is claimed to be determined by the appended claims.
Claims (5)
1. kind of computer based myeloplast automark method, which is characterized in that described to include the following steps:
Step 1: reading cell text information file, cell image CellImag is determined according to (x, y, R)；
Step 2: cell image CellImage is inputted preset disaggregated model, obtains the cell and belong to marrow difference leucocyte
Probability tables, the corresponding label class of maximum probability and its probability value are stored according to (x, y, R, label, Probablity) group
In the text structural information knitted；
Step 3: myeloplast classification figure layer image generates, such is belonged to according to the generation of the label label of default disaggregated model
The other all cells of same class are corresponded at it and use color at position (x, y, R) by other particular color ColorLabel
ColorLabel is marked with circle, generates classification figure layer image LabelLayerImage；
Step 4: the corresponding cell number of collect statistics label and all cell number mesh, generate different classes of myeloplast
Accounting information, numerical value visualization and carry out histogram visualization, by not in the cellular informatics of normal rates range class with red
Color box is marked.
2. computer based myeloplast automark method according to claim 1, it is characterised in that: the step
Preset disaggregated model carries out the model that classification may include different myeloplast types, prototype network structure to cell in rapid two
And leucocyte type can make corresponding change according to demand.
3. a kind of automatic tagging system of computer based myeloplast, special to be, the system comprises slide numbers
Change module, digitized image stripping and slicing module, cell image fixation and recognition module, slide cellular informatics Macro or mass analysis module and visual
Change module:
The slide digital module: the pathology of each actual patient is obtained to scan multiplying power ScanPlan by scanner and is cut
The digitized image DigitalImage of piece, digitized image is to scan specific format storage；
The digitized image stripping and slicing module: by reading digitized image DigitalImage, by digitized image
The fritter Blocks that DigitalImage is cut into 2048*2048 facilitates the subsequent progress parallel processing of computer to improve efficiency；
The cell image fixation and recognition module: Blocks is inputted into queue to be processed, computer is each time from queue to be processed
Automatic ProperNum image data of reading is automatically positioned dividing method by the cell of myeloplast and quickly identifies that marrow is white
Cell, and the myeloplast of identification is input in trained corresponding neural network model obtain its correspond to classification information and
Its probability, and by relevant information according to (x, y, R, label, Probablity) storage to CellsPosition message structure text
In this；
The slide cellular informatics Macro or mass analysis module: read CellsPosition message structure text, according to wherein (x, y,
R, label, Probablity) the corresponding cell number of collect statistics label and all cell number mesh, generate different classes of bone
The accounting information of marrow leucocyte；
The visualization model: numerical value visualization is carried out according to slide cellular informatics Macro or mass analysis module results and carries out histogram
Visualization, will not be marked in the cellular informatics of normal rates range class with red boxes；Simultaneously by classification figure layer image
LabelLayerImage and raw digitized image DigitalImage progress is compound, can check single classification according to selection
Myeloplast tag image either multi-class myeloplast tag image.
4. the automatic tagging system of computer based myeloplast according to claim 3, it is characterised in that: the glass
The common multiplying power of ScanPlan is 100 times of object lens in piece digital module.
5. the automatic tagging system of computer based myeloplast according to claim 3, it is characterised in that: described thin
The value of ProperNum is related with the preset disaggregated model complexity in claim 2 in born of the same parents' framing identification module.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910031626.1A CN109754398A (en) | 2019-01-14 | 2019-01-14 | A kind of computer based myeloplast automark method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910031626.1A CN109754398A (en) | 2019-01-14 | 2019-01-14 | A kind of computer based myeloplast automark method and system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109754398A true CN109754398A (en) | 2019-05-14 |
Family
ID=66404673
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910031626.1A Pending CN109754398A (en) | 2019-01-14 | 2019-01-14 | A kind of computer based myeloplast automark method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109754398A (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110728666A (en) * | 2019-10-08 | 2020-01-24 | 中山大学附属第三医院 | Typing method and system for chronic nasosinusitis based on digital pathological slide |
| CN111062346A (en) * | 2019-12-21 | 2020-04-24 | 电子科技大学 | Automatic leukocyte positioning detection and classification recognition system and method |
| CN111476754A (en) * | 2020-02-28 | 2020-07-31 | 中国人民解放军陆军军医大学第二附属医院 | Artificial intelligence auxiliary grading diagnosis system and method for bone marrow cell image |
| CN111667472A (en) * | 2020-06-08 | 2020-09-15 | 江西卫生职业学院 | Artificial intelligent analysis method for bone marrow cell morphology |
| WO2022021224A1 (en) * | 2020-07-30 | 2022-02-03 | 深圳迈瑞生物医疗电子股份有限公司 | Sample classification method and apparatus, and storage medium |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101088089A (en) * | 2004-02-23 | 2007-12-12 | 鹿特丹伊拉斯姆斯大学医疗中心 | Classification, diagnosis and prognosis of acute myeloid leukemia by gene expression profiling |
| CN103745210A (en) * | 2014-01-28 | 2014-04-23 | 爱威科技股份有限公司 | Method and device for classifying white blood cells |
| CN105228749A (en) * | 2013-03-13 | 2016-01-06 | 塔霍农村卫生研究所有限公司 | Portable blood cell count monitor |
| US20170132450A1 (en) * | 2014-06-16 | 2017-05-11 | Siemens Healthcare Diagnostics Inc. | Analyzing Digital Holographic Microscopy Data for Hematology Applications |
| CN107729932A (en) * | 2017-10-10 | 2018-02-23 | 李强 | Bone marrow cell labeling method and system |
| CN108257124A (en) * | 2018-01-23 | 2018-07-06 | 江苏康尚生物医疗科技有限公司 | A kind of white blood cell count(WBC) method and system based on image |
-
2019
- 2019-01-14 CN CN201910031626.1A patent/CN109754398A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101088089A (en) * | 2004-02-23 | 2007-12-12 | 鹿特丹伊拉斯姆斯大学医疗中心 | Classification, diagnosis and prognosis of acute myeloid leukemia by gene expression profiling |
| CN105228749A (en) * | 2013-03-13 | 2016-01-06 | 塔霍农村卫生研究所有限公司 | Portable blood cell count monitor |
| CN103745210A (en) * | 2014-01-28 | 2014-04-23 | 爱威科技股份有限公司 | Method and device for classifying white blood cells |
| US20170132450A1 (en) * | 2014-06-16 | 2017-05-11 | Siemens Healthcare Diagnostics Inc. | Analyzing Digital Holographic Microscopy Data for Hematology Applications |
| CN107729932A (en) * | 2017-10-10 | 2018-02-23 | 李强 | Bone marrow cell labeling method and system |
| CN108257124A (en) * | 2018-01-23 | 2018-07-06 | 江苏康尚生物医疗科技有限公司 | A kind of white blood cell count(WBC) method and system based on image |
Non-Patent Citations (1)
| Title |
|---|
| 陈颖等: "数字化全玻片助力人工智能病理图像决策", 《海军军医大学学报》 * |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110728666A (en) * | 2019-10-08 | 2020-01-24 | 中山大学附属第三医院 | Typing method and system for chronic nasosinusitis based on digital pathological slide |
| CN111062346A (en) * | 2019-12-21 | 2020-04-24 | 电子科技大学 | Automatic leukocyte positioning detection and classification recognition system and method |
| CN111476754A (en) * | 2020-02-28 | 2020-07-31 | 中国人民解放军陆军军医大学第二附属医院 | Artificial intelligence auxiliary grading diagnosis system and method for bone marrow cell image |
| CN111476754B (en) * | 2020-02-28 | 2022-12-09 | 中国人民解放军陆军军医大学第二附属医院 | Bone marrow cell image artificial intelligence auxiliary grading diagnosis system and method |
| CN111667472A (en) * | 2020-06-08 | 2020-09-15 | 江西卫生职业学院 | Artificial intelligent analysis method for bone marrow cell morphology |
| WO2022021224A1 (en) * | 2020-07-30 | 2022-02-03 | 深圳迈瑞生物医疗电子股份有限公司 | Sample classification method and apparatus, and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109754398A (en) | A kind of computer based myeloplast automark method and system | |
| CN112508850B (en) | Deep learning-based method for detecting malignant area of thyroid cell pathological section | |
| Gallardo-Caballero et al. | Precise pollen grain detection in bright field microscopy using deep learning techniques | |
| CN111079620B (en) | White blood cell image detection and identification model construction method and application based on transfer learning | |
| CN111738083B (en) | Training method and device for face recognition model | |
| CN108596046A (en) | A kind of cell detection method of counting and system based on deep learning | |
| CN114219786B (en) | Chromosome karyotype analysis method and system based on deep learning | |
| Bergler et al. | FIN-PRINT a fully-automated multi-stage deep-learning-based framework for the individual recognition of killer whales | |
| CN112365471A (en) | Cervical cancer cell intelligent detection method based on deep learning | |
| Brenner et al. | Automated classification of normal and abnormal leukocytes | |
| CN109948429A (en) | Image analysis method, device, electronic equipment and computer-readable medium | |
| CN115170518A (en) | Cell detection method and system based on deep learning and machine vision | |
| CN113160185A (en) | Method for guiding cervical cell segmentation by using generated boundary position | |
| Yang et al. | Pathminer: a web-based tool for computer-assisted diagnostics in pathology | |
| CN117252813B (en) | Cervical liquid-based cell detection and identification method and system based on deep learning | |
| CN113096071A (en) | CHS-Net network-based medical image segmentation method | |
| CN120766940A (en) | An intelligent diagnosis system for cervical cell images based on a multimodal visual language model | |
| CN118629037A (en) | An immune cell image segmentation method and system based on multi-instance learning | |
| CN106503747A (en) | A kind of image recognition statistical analysis system | |
| CN104966109A (en) | Image classification method and device for medical laboratory test sheets | |
| CN115359325B (en) | Training method, device, equipment and medium for image recognition model | |
| CN113076960A (en) | Image classification method and device based on multi-scale feature iterative fusion network | |
| CN114898866B (en) | Thyroid cell auxiliary diagnosis method, thyroid cell auxiliary diagnosis equipment and thyroid cell auxiliary diagnosis storage medium | |
| CN113963199A (en) | A medical waste identification method based on multiple sensor feature fusion and machine learning | |
| CN113409293A (en) | Pathology image automatic segmentation system based on deep learning |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190514 |