CN110085315A - Cytomorphology artificial intelligence diagnoses new method automatically - Google Patents
Cytomorphology artificial intelligence diagnoses new method automatically Download PDFInfo
- Publication number
- CN110085315A CN110085315A CN201910301730.8A CN201910301730A CN110085315A CN 110085315 A CN110085315 A CN 110085315A CN 201910301730 A CN201910301730 A CN 201910301730A CN 110085315 A CN110085315 A CN 110085315A
- Authority
- CN
- China
- Prior art keywords
- artificial intelligence
- intelligence
- cytomorphology
- cell
- human brain
- 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
Classifications
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
- G16H50/20—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for computer-aided diagnosis, e.g. based on medical expert systems
Landscapes
- Health & Medical Sciences (AREA)
- Engineering & Computer Science (AREA)
- Biomedical Technology (AREA)
- Medical Informatics (AREA)
- Public Health (AREA)
- Data Mining & Analysis (AREA)
- Databases & Information Systems (AREA)
- Pathology (AREA)
- Epidemiology (AREA)
- General Health & Medical Sciences (AREA)
- Primary Health Care (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The present invention relates to a kind of cytomorphology artificial intelligence to diagnose new method automatically, belongs to medical field.The present invention is directly to establish the mathematical model of " the final rule " of expert's grade " human brain intelligence ", as the mathematical model and algorithm of " artificial intelligence ", without by learning process early period and establishing huge database.The present invention has further pushed the development of artificial intelligence negatoscopy, is the technological break-through again in the field, has broad application prospects.
Description
Technical field
The present invention relates to a kind of cytomorphology artificial intelligence to diagnose new method automatically, belongs to medical field.
Background technique
Artificial intelligence is a branch of computer science, it attempts to understand essence of intelligence, and is produced a kind of new
The intelligence machine that can be made a response in such a way that human intelligence is similar, the research in the field include robot, language identification, image
Identification, natural language processing and expert system etc..Artificial intelligence is since the birth, and theory and technology is increasingly mature, application field
Also constantly expand, it is contemplated that the following artificial intelligence bring sci-tech product, it will be the wisdom of humanity " container ".Artificial intelligence
Can consciousness to people, thinking information process simulation.Artificial intelligence is not the intelligence of people, but can think deeply as people,
It may also be more than the intelligence of people.
The fast development of artificial intelligence technology keeps Cytomorphology more efficient and automation, meanwhile, because it is not tired
The unlimited dilatancy of Lao Xing, the accuracy of analysis and database greatly improve the accuracy of diagnosis.What is certain is that
Artificial intelligence technology will necessarily make morphological diagnosis reach a new technology height.
From the substantial angle of artificial intelligence concept, main target is the work essence for exploring and imitating first human brain, so
It is optimized again according to the characteristics of computer equipment afterwards, is finally reached intelligentized target.Therefore, artificial intelligence application is in cell
The technological essence of Morphologic Diagnosis is the human brain diagnostic mode for simulating experienced read tablet doctor, and important foundation is using meter
Calculation machine technology constructs " read tablet experience ".
But the so-called artificial intelligence technology in the field at present, most of is not real artificial intelligence.These technologies,
Image data base substantially is constructed by a large amount of positive piece, then by after the cell smear digitlization for needing to diagnose, is used tricks
Calculation machine software is compared with the content of database, chooses immediate comparison result finally to generate diagnosis report.This skill
Art, essence are exactly simple " image comparison software ", are strongly depend on the scale of image data base, theoretically, database
Huger, diagnosis is higher.It is exactly because technical threshold is low, so that a large amount of enterprises and individuals pour in the field, and constantly heat
Inner feelings is in the positive piece of collection everywhere and establishes image data base as perfect as possible.
What is certain is that achieving the purpose that diagnose automatically using the image comparison technology of low side, much can't resolve
Actual read tablet problem.To find out its cause, being because cellular morphology is ever-changing, ethnic group is different, region is different, the age is different, a
Body sample time difference etc., determines that cytomorphology is at all impossible to exhaust.It cannot be exhaustive, it is meant that can not receive
All possible database is received, the database of foundation is also impossible to the possibility of all positives is all included.Therefore, this
Kind relies on the image comparison technology of database, naturally can be in the state for having blind area forever.
Therefore, it is necessary to find the read tablet mode of " human brain intelligence ", computer system could be used to be imitated and developed.
I.e., it is necessary to find " the human brain experience " for fitting using computer mould and rebuilding experienced read tablet doctor, be only real artificial intelligence
Energy technology, also need nots rely upon and establishes huge database.
Summary of the invention
Inventor has summed up principle and working method that " human brain intelligence " carries out cellular morphology diagnosis through numerous studies,
And the mathematical model of " human brain intelligence " has been successfully established it.
It is obvious that experienced read tablet doctor, it is impossible to store a large amount of clearly images in the brain of oneself, this is not
The working method of human brain.Why intelligence is known as, necessarily because cell shape is gradually grasped and summarized to human brain by study
The regular method that the feature of state and comprehensive various aspects morphological feature are judged.This is only the working method of " human brain intelligence ".
That is, the working method of " human brain intelligence ", does not rely on mass image data and is compared, but in various typical cells
It in morphological feature, constantly summarizes and refines regular content, and in diagnostic work in the future, continue to optimize and adjust the regularity
Content, having eventually formed can be by true " the final rule " solved repeatedly.
Since inventor itself is exactly veteran cell read tablet person, the rule and read tablet stream to Cytomorphology
Thought process in journey has sufficient subjective feeling.Therefore, inventor determines to start with from the thought process of human brain read tablet, in conjunction with
The personnel training process of cyto-diagnosis, that is, how to make an inexperienced new person grow into can skilled read tablet person, and then sufficiently
The working principle of analysis " human brain intelligence ", it is final to lay the foundation to establish artificial intelligence using computer.
Actual personnel training process proves, to allow beginner to improve read tablet level rapidly, it is big to be not meant to the beginner
Read tablet is measured, but reads typical positive piece more, and synchronous training diagnosis rule, it is made to grasp read tablet principle and from typical positive piece
Diagnosis rule is summarized rapidly.That is, beginner is not allowed for remember great amount of images, but remember the form rule of typical cells
Rule.As a result, it has been recognised by the inventors that the performance of " human brain intelligence ", relies primarily on two aspect bases.First is that the mode of thinking of read tablet person,
The natures factor such as personality, is equivalent to mathematical model.Same training, different people's progress speed are different.This be mainly because
Caused by for the difference of nature and learning method, i.e. the building of mathematical model is different.It on the other hand is exactly to sum up
Whether diagnosis rule, which meets diagnosis, requires, and is equivalent to algorithm.
Therefore, the emphasis of invention has been placed on and how to have established expert's grade mathematical model and expert grade two side of algorithm by inventor
Face.As long as directly establishing the above two aspects content, " artificial intelligence " of the invention is not necessarily to large scale database, Ji Keshi at all
Existing high-caliber automatic diagnosis.
Current " image comparison " technology, it is necessary to first solve image procossing (emphasis is that image sharpening is strengthened), establish form
Database solves a series of problems, such as image compares.But inventor it has been investigated that, " human brain intelligence " or " artificial intelligence " is not required to
Want such process.
Firstly, the step for not needing image procossing.For current image scanning techniques, the fineness that no matter scans
Or the reduction degree of color and image, is all sufficient for the requirement of image digitazation.And " human brain intelligence " does not need big spirogram
Piece is supported, therefore, it has been recognised by the inventors that " human brain intelligence " or the core of " artificial intelligence " are to carry out figure using " human brain experience "
The step for picture analysis, image procossing, not should belong to core technology scope.In addition, the work premise of " human brain intelligence ", is just answered
This is for qualified, cell smear clearly, excellent, without being built upon on the basis of traditional low quality scraping blade.And mesh
Preceding cervical liquid-based cells technology also includes the non-gynecological liquid-based technology developed in recent years, and mission is exactly to produce height
The cytologic slide of quality.Therefore, it has been recognised by the inventors that the important process premise and basis of " human brain intelligence " or " artificial intelligence ",
Should be built upon on outstanding cytologic slide since piece it is outstanding (plating cells uniformly, form complete display, nothing
Impurity interference, staining versus's degree are high), image processing problem just is considered without emphasis at all naturally.
Secondly, " artificial intelligence " is to carry out practical read tablet application so that " human brain intelligence " foundation " final rule " is result
, therefore, it has been recognised by the inventors that only needing to be successfully established the mathematical model or the identical function of building of " the final rule " of " human brain intelligence "
The software architecture of energy, can be realized dynamic diagnosis.The complexity of " artificial intelligence " can thus be greatly simplified.
Finally, inventor is through numerous studies, discovery " human brain intelligence " is actually parallel computation, is reflected in macroscopically, just
It is that there is " affective thinking ".Rather than the logical thinking of " image comparison technology ".The mode of parallel computation, exactly when read tablet person sees
It is the Synchronization Analysis for carrying out indices when to cell, the point of penetration of analysis is simultaneously not fixed, this will be according to personality, immediately impression
Etc. indexs carry out the personalisation process of adaptation to local conditions.Therefore, it has been recognised by the inventors that the mathematical model of " artificial intelligence ", it should be
The set of multiple mathematical models of parallel computation, and calculating process not have strong logicality, as long as being divided into main indicator
With secondary index.Such " artificial intelligence ", the form of expression will be more closely similar to the perception expression of " human brain intelligence ".
To sum up, the cytomorphology artificial intelligence of inventor's invention diagnoses new method automatically, is different from following three
At present the characteristics of technology:
1) it is not necessarily to huge database.Typical morphological data need to only be collected.
2) mathematical model and itself, so that it may be directly realized by automatic diagnosis.Without a large amount of learning process early period.Because of itself
It is exactly expert level, therefore, " artificial intelligence " of the invention, it might even be possible to directly carry out high level read tablet with human expert and grind
It begs for.
3) " artificial intelligence " of the invention does not consider the quality problems of cell smear, defaults aiming at the high-quality of qualification
Piece is measured, conventional wiper blades should not be taken as assessment basis and standard.Because even being current liquid-based technology, can also accomplish excellent
Elegant quality this point.
Main contents of the invention are exactly the mathematical modulo for directly establishing " the final rule " of " human brain intelligence " of expert's grade
Type, using this mathematical model as the mathematical model of " artificial intelligence " and algorithm, without by learning process early period and huge number
According to library." the artificial intelligence of " final rule " directly founding mathematical models and algorithm of all " human brain intelligence " by mature read tablet person
Can " cytomorphology automatic diagnosis method, belong to interest field of the invention.
The mathematical model and algorithm that " the final rule " of " human brain intelligence " that inventor proposes in embodiment is established, are all
One in more specific solutions.It is obvious that the specific solution of the model will be not limited to embodiment.
The present invention has further pushed the development of artificial intelligence negatoscopy, is the technological break-through again in the field, tool
Have broad application prospects.
Embodiment
Inventor proposes the scheme of being implemented as follows according to mature read tablet experience and numerous studies.This mathematical model and calculation
Method can directly carry out the automatic diagosis of cytomorphology, establish database without early period.
Since the present invention does not depend on huge database, the main body of this artificial intelligent software system is the mould of " human brain intelligence "
Quasi-, therefore, this software systems overcomes " hardware system configuration is high, data storage is huge " of other current artificial intelligence softwares etc.
Deficiency, the present invention is of less demanding to hardware configuration instead, and is more focused on the stability of system.This is just universalness of the invention
Good hardware foundation has been laid with cost effective.
The present embodiment is the thinking implemented for demonstration and own characteristic of the invention, belongs to basic research scheme, very bright
Aobvious, actual application scheme wants increasingly complex.
The hardware system of the present embodiment, is divided into three parts.
First part is the operation hardware system of artificial intelligence software.Using the tower two-way of Dell PowerEdge T440
Server.Dual processors, CPU E5-2600V3, memory 2T, 4 pieces of 3.5 cun of hard disks, total capacity 8T.Cost is not high, and technology maturation is adopted
Purchase approach is abundant, easy maintenance, the operation is stable.Second part is digital slices automatic scanning system.Using Leca
Leica SCN400 integration slide scanning system, the cell image information collection source as high quality.The system has already passed through
A large amount of clinical users are examined, mature and stable, are also easy buying.Part III is exactly the cytologic slide of high-quality.The present embodiment
Sample process and film-making are carried out using liquid based cytology technology.But when dyeing, the time of nucleus and cytoplasm and artificial reading
Piece is slightly different.Due to being automatic collection image, to increase image recognition rate, the present embodiment specially suitably adds dyeing time
It is long, the effect slightly contaminated deeply is formed, to increase core paddle contrast when chromoscan.
The software architecture of the present embodiment uses C language.The langue maturity is high, has a wide range of application, hardware supported
Range is big, and versatility is good.
Emphasis of the invention is the mathematical model and algorithm for establishing " human brain intelligence ", and therefore, the present embodiment emphasis is simulation
The thinking decision mode of " human brain intelligence " and the key algorithm for generating conclusion.
The present embodiment is the basic software systems of demonstration, and therefore, emphasis is the basic operation thinking and mode of demonstration.
This algorithm itself is that basic diagnostic element has been carried out mathematical model, it is evident that when practical application, the number that needs to establish
It learns model and is necessarily much more complicated than the present embodiment.
The algorithm of the present embodiment is divided into three parts, is three mathematical models respectively, three bases diagnosed with expression cell
Plinth parameter.This three basic parameters are " cell category ", " karyoplasmic ratio " and " cell nuclear state " respectively.Certainly, it
At accurate diagnosis, it is also necessary to a lot of other data, but this three data be it is basic, as basis of software function presentation
It is described the problem enough.
Three mathematical models are respectively with 1. number equation, 2. number equation and 3. number equation is expressed.In three equations
Hold as follows.
1. number equation is used to determine " karyoplasmic ratio ".Wherein, XmxsIt is the numerical value of basic core paddle ratio, is under normal circumstances
Core paddle ratio, as measure overall core paddle than whether normal reference value.Because of cell volume affirmative, special circumstances bigger than nucleus
It is bare nucleus or cell membrane almost close to nucleus, therefore, which is at least 1.Each cell category, the value is all different, can basis
Expert's grade diagnostic experiences or corresponding diagnostic criteria, input the basic core paddle ratio of each cell types in advance.K is adjustment system
Number is used to balance error.It is subordinating degree function in braces.Through this equation, each type directly relatively accurately can be quickly determined
The core paddle ratio of cell, and first come out the cell screening of extremely abnormal numerical value.
2. equation be used to determine " cell nuclear state ", specifically, party's formula can determine " nuclear membrane smooth degree " and
" the regular degree of chromosome " (being commonly referred to as " kernel state ").Nuclear membrane edge, it can be understood as be one continuous irregular
Curve is made of the connection of many mini line segments.The curvature of each mini line segment then represents the smooth journey of this section of slight arc
Degree.The curvature of each slight arc section is summarized again, comprehensive assessment goes out whole smoothness.It is obvious that final result is smaller, then light
Slippage degree is higher.The cell of each type, the smooth degree of nucleus is different, can be according to expert's grade diagnostic experiences or corresponding
Diagnostic criteria, input the nucleus state value of each cell types in advance, as assessment foundation.In this equation, round bracket
It is interior, it is the distance at the head and the tail both ends of each slight arc section, and equation is whole, then is that each slight arc section is smooth
The overall summing function of degree, then its square root is taken, it can relatively accurately show the smooth degree of the nucleus.The letter of this equation
It is single to be illustrated, it can quickly filter out the especially abnormal cell of cell nuclear state.
3. equation is used to determine " cell category ".Because mainly observing the section of cell under microscope, we
Formula is measured using the scale of two-dimensional space, without considering three-dimensional dimension.This algorithm, broken it is usual with cell dia size come
The method of measure of cell type, and area size is used instead to measure.Mainly in view of having cell folding during practical film-making
It is folded or just it is seen that pne cell, these situations, if just will appear the very extreme situation of number using diameter data,
It is unfavorable for interpretation and analysis.It, can be to avoid above situation if measured using area data.In this equation, two brackets
The cell dia in two orthogonal directions of interior expression, product are the cell area.Meanwhile it is if vertical each other
The numerical value of both direction have big difference, it was demonstrated that be pne cell or folding, then abandon the cell interpretation.The right side of first bracket
The negative one power of the S at upper angle, plays deviation-correcting function.S value is the basic karyoplasmic ratio of the cell types.Party's formula can quickly be sentenced
Disconnected cell category, moreover it is possible to realize the synchronous cell for excluding abnormal posture, it is high-efficient, it is more conform with the read tablet side of " human brain intelligence "
Formula.
" human brain intelligence " is actually not sentenced according to fixed logical order in this three basic datas of application
It is other, but comprehensive descision.For example, right mind process, is determining first " cell category ", then whether " karyoplasmic ratio " is observed
The normal level for meeting the cell types further judges " nucleus shape finally, for the situation of " karyoplasmic ratio " exception
State ", if whether nuclear membrane is smooth, whether kernel contaminates deeply.But in fact, being likely to just see " cell nuclear state " in the first step
Then abnormal cell is observed " karyoplasmic ratio " again, last resort " cell category " judges whether to meet the cell types
Normal index.Above it is not difficult to find out that, actual diagnosis process, not instead of logical process, random synchronizing process.But no matter first
Afterwards, three indexs all need to refer to.
Therefore, this three equations are in the present embodiment software running process, no fixed logic sequence, and three equations
Formula is synchronous operation.This has also fully demonstrated the essential distinction of " artificial intelligence " and " image comparison ".
The implementation process of the present embodiment is as follows；
Firstly, with C language framework by 3 equations and mating program construction " artificial intelligence " software." the people of the present embodiment
Work intelligence " software, in addition to main program, at least there are also " all types of cell sample positive data library " and " slide scanning system behaviour
Make interface " etc. modules.Then, with slide scanning system, the qualified liquid-based smear of various types of cells sample is collected, number is made and cuts
Piece, and typical positive cellular morphology is classified and is stored in the taxonomy database of this " artificial intelligence " software.It is emphasized that this
Embodiment is not necessarily to mass data, every kind of cell sample, it is only necessary to collect the morphological data of 50 typical positive cells.Most
Afterwards, slide scanning system is accessed to the host of the present embodiment." artificial intelligence " that a set of demonstration can be set up above is examined automatically
Disconnected system.
The workflow of this " artificial intelligence " auto-check system is as follows:
After armamentarium booting, each section completes self-test, into working condition.Slide to be diagnosed is put into one by operator
Change in slide scanning system, and select cell sample classification in the operation interface of this system, at this point, this system and integrated glass
Piece scanning system software auto-associating, using integrated slide scanning system software as itself sub- software, the two completes function
Connection.Operator selects after starting diagnostic function, and integrated slide scanning system starts to carry out the work of slide digital collection,
Collected cellular morphology information is automatically imported into this system.But this system be not by the way of " image comparison " software,
It is not that each cell is scanned one time one by one to be compared, such efficiency is too low.This system is the work using " human brain intelligence "
Mode.This system is after whole cells wait diagnose plectrum all complete digitlization, using three equations, by slide information
Interior cellular morphology information carries out disposable outline assessment.First filter out 1. number equation (" karyoplasmic ratio ") and 2. number equation
The especially abnormal cellular morphology of index of (" cell nuclear state "), it is synchronous with based on 3. number equation (" cell category ")
With reference to carrying out Synchronous Screening.If occurring without especially abnormal index in the assessment of this outline, being then considered as negative sample.If general
Exception is slightly evaluated, then to the abnormal cell filtered out, carries out the joint emphasis assessment of three equations again, is such as still different
Often, then diagnosis report is provided according to the information of database.If second of emphasis assessment eliminates exception, then it is considered as negative sample.
From first time Pre-Evaluation to later assessment each time, until issuing diagnosis report, this system will all be automatically recorded, in order to people
Work audit.The automatic diagnostic process of the present embodiment, compared to other technologies, efficiency at least improves 50 times.
From embodiment, it is not difficult to find out that, the present invention is entirely the working method of simulation " human brain intelligence ".Its advantage is very prominent,
Do not depend on that large database concept, can not network independent operating, hardware requirement are lower to be easy to universal and reduce cost such as.
The present invention relative to current technology, advantage be it is fairly obvious, also represent the important of " artificial intelligence " technology
Developing direction.
Claims (4)
1. a kind of cytomorphology artificial intelligence diagnoses new method automatically, which is characterized in that it is not necessarily to learning process early period, it can be direct
Carry out expert's grade diagnosis.
2. a kind of cytomorphology artificial intelligence diagnoses new method automatically, which is characterized in that without establishing huge database.
3. a kind of cytomorphology artificial intelligence diagnoses new method automatically, which is characterized in that directly establish " the human brain intelligence of expert's grade
The mathematical model of " the final rule " of energy ", mathematical model and algorithm as " artificial intelligence ".
4. a kind of cytomorphology artificial intelligence diagnoses new method automatically, which is characterized in that only need to collect a small amount of typical sun
Property cellular morphology data.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910301730.8A CN110085315A (en) | 2019-04-15 | 2019-04-15 | Cytomorphology artificial intelligence diagnoses new method automatically |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910301730.8A CN110085315A (en) | 2019-04-15 | 2019-04-15 | Cytomorphology artificial intelligence diagnoses new method automatically |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110085315A true CN110085315A (en) | 2019-08-02 |
Family
ID=67415140
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910301730.8A Pending CN110085315A (en) | 2019-04-15 | 2019-04-15 | Cytomorphology artificial intelligence diagnoses new method automatically |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110085315A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110634134A (en) * | 2019-09-04 | 2019-12-31 | 杭州憶盛医疗科技有限公司 | Novel artificial intelligent automatic diagnosis method for cell morphology |
| CN110767307A (en) * | 2019-09-20 | 2020-02-07 | 杭州憶盛医疗科技有限公司 | Novel artificial intelligent automatic diagnosis method for cell morphology |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106778037A (en) * | 2017-01-12 | 2017-05-31 | 武汉兰丁医学高科技有限公司 | A kind of analysis method based on the diagnosis of cell image Cloud Server |
| CN108288506A (en) * | 2018-01-23 | 2018-07-17 | 雨声智能科技（上海）有限公司 | A kind of cancer pathology aided diagnosis method based on artificial intelligence technology |
-
2019
- 2019-04-15 CN CN201910301730.8A patent/CN110085315A/en active Pending
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106778037A (en) * | 2017-01-12 | 2017-05-31 | 武汉兰丁医学高科技有限公司 | A kind of analysis method based on the diagnosis of cell image Cloud Server |
| CN108288506A (en) * | 2018-01-23 | 2018-07-17 | 雨声智能科技（上海）有限公司 | A kind of cancer pathology aided diagnosis method based on artificial intelligence technology |
Non-Patent Citations (2)
| Title |
|---|
| 张缨等: "图像处理和人工神经网络在肺癌细胞病理诊断中的应用", 《中华胸心血管外科杂志》 * |
| 龚平等: "乳腺癌针吸细胞形态定量的人工神经网络诊断模型的建立及应用研究", 《中国体视学与图像分析》 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110634134A (en) * | 2019-09-04 | 2019-12-31 | 杭州憶盛医疗科技有限公司 | Novel artificial intelligent automatic diagnosis method for cell morphology |
| CN110767307A (en) * | 2019-09-20 | 2020-02-07 | 杭州憶盛医疗科技有限公司 | Novel artificial intelligent automatic diagnosis method for cell morphology |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106779087B (en) | A kind of general-purpose machinery learning data analysis platform | |
| JP2022529557A (en) | Medical image segmentation methods, medical image segmentation devices, electronic devices and computer programs | |
| CN102567734B (en) | Specific value based retina thin blood vessel segmentation method | |
| CN109376636A (en) | Eye ground image classification method based on capsule network | |
| CN107342810A (en) | Deep learning Brilliant Eyes figure analysis method based on convolutional neural networks | |
| CN104346481B (en) | A kind of community detection method based on dynamic synchronization model | |
| CN105808665B (en) | A kind of new image search method based on cartographical sketching | |
| CN108256482A (en) | A kind of face age estimation method that Distributed learning is carried out based on convolutional neural networks | |
| CN109544507A (en) | A kind of pathological image processing method and system, equipment, storage medium | |
| CN110085315A (en) | Cytomorphology artificial intelligence diagnoses new method automatically | |
| CN106408576B (en) | Automatic region of interest segmentation method and system based on three-dimensional ultrasonic image | |
| CN108734138A (en) | A kind of melanoma skin disease image classification method based on integrated study | |
| CN111767952B (en) | Interpretable lung nodule benign and malignant classification method | |
| CN109102498A (en) | A kind of method of cluster type nucleus segmentation in cervical smear image | |
| CN109584153A (en) | Modify the methods, devices and systems of eye | |
| CN112365471B (en) | Cervical cancer cell intelligent detection method based on deep learning | |
| CN108288048A (en) | Based on the facial emotions identification feature selection method for improving brainstorming optimization algorithm | |
| CN110377659A (en) | A kind of intelligence chart recommender system and method | |
| CN110288097A (en) | A kind of method and relevant apparatus of model training | |
| CN111062260A (en) | Automatic generation method of facial cosmetic recommendation scheme | |
| CN107506362A (en) | Image classification based on customer group optimization imitates brain storage method | |
| CN109697459A (en) | One kind is towards optical coherence tomography image patch Morphology observation method | |
| CN106599529A (en) | Digital slice processing system | |
| Zhang et al. | Abnormal region detection in cervical smear images based on fully convolutional network | |
| Lin et al. | Fedcluster: A federated learning framework for cross-device private ecg classification |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| CB02 | Change of applicant information | ||
| CB02 | Change of applicant information |
Address after: Qinhuai Yang Gong Jing District of Nanjing City, Jiangsu Province, No. 34 34 210002 Applicant after: Qinhuai medical area, Eastern theater General Hospital Address before: Qinhuai Yang Gong Jing District of Nanjing City, Jiangsu Province, No. 34 34 210002 Applicant before: NO.81 HOSPITAL OF PLA |
|
| WD01 | Invention patent application deemed withdrawn after publication | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20190802 |