CN108805101A - A kind of recognition methods of the parasite egg based on deep learning - Google Patents
A kind of recognition methods of the parasite egg based on deep learning Download PDFInfo
- Publication number
- CN108805101A CN108805101A CN201810688609.0A CN201810688609A CN108805101A CN 108805101 A CN108805101 A CN 108805101A CN 201810688609 A CN201810688609 A CN 201810688609A CN 108805101 A CN108805101 A CN 108805101A
- Authority
- CN
- China
- Prior art keywords
- picture
- parasite egg
- deep learning
- worm
- displaing micro
- 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
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/60—Type of objects
- G06V20/69—Microscopic objects, e.g. biological cells or cellular parts
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/60—Type of objects
- G06V20/69—Microscopic objects, e.g. biological cells or cellular parts
- G06V20/695—Preprocessing, e.g. image segmentation
Abstract
The invention discloses a kind of recognition methods of the parasite egg based on deep learning, and described method includes following steps：S1：Parasite egg picture is zoomed into displaing micro picture, displaing micro picture is identified, and clears up the picture for not including worm's ovum, after cleaning does not include the picture of worm's ovum, the picture after cleaning is chosen；S2：The displaing micro picture of selection is established into sample database, parasite egg type and position in sample database displaing micro picture is marked, is learnt automatically from displaing micro picture sample database using artificial intelligence deep learning model and extract characteristic information；S3：Using the parasitic ovum type marked in step S2, the characteristic value of parasite egg is extracted, property data base is established, the test samples library discrimination in property data base, and recognition result is exported, and according to discrimination result, the sample database picture for failing identification is rejected, new samples library is formed.
Description
Technical field
The present invention relates to a kind of recognition methods, and in particular to a kind of identification side of the parasite egg based on deep learning
Method.
Background technology
Parasitic disease is global public health security problem, and worm's ovum microscopy is crucial Prevention Technique, in clinic
It is widely used.It is identified however, worm's ovum microscopy can only rely on artificial naked eyes for a long time.Its method is cumbersome, and identification error
It is different with the experience and state of reviewer, lack objectivity, accuracy and stability.So using the higher identification of accuracy rate
Technology, indirect labor's microscopy will effectively improve many deficiencies of manual identified.
In recent years, many researchers are carried out parasite egg identification and are ground using the method for traditional artificial extraction characteristic value
Study carefully, but because method is limited to, causes discrimination that cannot meet adjuvant clinical application.Such as, auspicious dimension in nineteen ninety-five hole etc. has been carried out aobvious
The research of worm egg identifying system under micro mirror, correct recognition rata is close to 92%；Castanon in 2007 etc. utilizes Bayes's classification
Device realizes the identification for infecting poultry parasite, and discrimination reaches 85%.2016,11 kinds of main human's parasitisms such as Shen Haimo
The digitized description of worm worm's ovum and automatic identification research, discrimination 91.83%.Human bodies of the Hu little Fang based on Hadoop in 2016
The research and realization of parasite egg identifying system, discrimination 93%.However these researchs are often all based on more satisfactoryization
In the state of carry out, experimental method difficulty is competent at actual detection needs, especially in the case where impurity is more, it is difficult to obtain
Stable recognition result, while recognition efficiency is not high.
These study the actual conditions not considered：1. there are more impurity or the back ofs the body for the tested altimetric image in practical application
Scape is complicated.2. the type of parasite is more, form varies in color, and is difficult to find to be suitble to own if carrying out image preprocessing
The method of worm's ovum extracts characteristic value.3. image capturing device has differences, shooting environmental is different, even same worm
Ovum, same time, the image acquired in different photographers can also have larger difference.4. parasite egg itself is when different
Phase also has different forms, and some difference are very big.Above four kinds of situations, if using the side of traditional artificial extraction characteristic value
Method, can not obtain complete characteristic value data, and discrimination cannot be satisfied the needs of adjuvant clinical diagnosis at all.
Recognition methods based on deep learning can solve, for convolutional neural networks in the application of field of image recognition and
On the one hand speech needs to choose the algorithm for being suitble to image recognition, then needs to pay close attention in different classes of varying environment on the other hand and produce
The feature of raw image itself.
Invention content
The discrimination being identified the technical problem to be solved by the present invention is to existing method is low, is carried out using manual type
Identification needs to consume prodigious manpower and materials, and it is an object of the present invention to provide a kind of identification side of the parasite egg based on deep learning again
Method solves the problem above-mentioned.
The present invention is achieved through the following technical solutions：
A kind of recognition methods of the parasite egg based on deep learning, described method includes following steps：S1：It will be parasitic
Worm worm's ovum picture zooms into displaing micro picture, identifies displaing micro picture, and clears up the picture for not including worm's ovum, is not wrapped in cleaning
After picture containing worm's ovum, the picture after cleaning is chosen；S2：The displaing micro picture of selection is established into sample database, marks sample
Parasite egg type in this library displaing micro picture and position, using artificial intelligence deep learning model from displaing micro picture sample database
In learn automatically and extract characteristic information；S3：Using the parasitic ovum type marked in step S2, parasite egg is extracted
Characteristic value, establish property data base, the test samples library discrimination in property data base, and recognition result is exported, and root
According to discrimination as a result, rejecting the sample database picture for failing identification, new samples library is formed.S4：By the new samples library in step S3, profit
With neural network algorithm, optimization extraction characteristic value forms new feature Value Data library and examines new sample using new feature Value Data library
This library discrimination, will export inspection result.Currently, there are many kinds of methods for the research of parasite egg identification, all it is much real
The mode of room operation is tested, this kind of laboratory research is often all based under more idealistic state and carries out, and experimental method is difficult
Competent actual detection needs, especially in the case where impurity is more, it is difficult to obtain stable recognition result, while identify effect
Rate is not high.When hospital is studied or carries out clinical diagnosis, need to carry out Classification and Identification to parasite egg, existing
Under method, the recognition methods recognition efficiency of use is low, and using full manual method, many hospitals can not undertake again, therefore this Shen
Please file under the premise of effectively improving recognition methods, reduce the utilization of professional to the greatest extent, wanting for discrimination can either be met
It asks, and cost is reduced for hospital.
First, worm's ovum picture is amplified under the microscope, amplification becomes displaing micro picture, and enlargement ratio at least needs
400 times, because of the displaing micro picture under the microscope of 40X10 or more, worm's ovum feature is apparent in the visual field, and dry without other impurity
It disturbs, if less than 400 times, amplification effect is bad, it is easy to occur many impurity and extra worm's ovum in the visual field, be unfavorable for figure
Piece is analyzed.By after selection i.e. clear up after displaing micro picture establish sample database, in sample database using convolutional neural networks algorithm into
Then the extraction and optimization of row characteristic value form a characteristic value data library, when higher than one setting of result discrimination of identification
Threshold value when, be preferably here 95%, and discrimination is excessively high, and the consumption of energy consumption can be very big, if discrimination is less than 90%,
It is unfavorable for the actual use of hospital, what preferably discrimination 95% can be best takes into account energy consumption and actual use.When discrimination is higher than
Threshold value, if then successfully the image after identification while exporting primary result, is established one by output as a result, less than this is preset
New samples picture library carries out characteristics extraction with after optimization by convolutional neural networks algorithm secondary, forms a new feature value number
According to library, result is exported.It when threshold value is consistently lower than threshold value, repeats the above steps always, until discrimination reaches threshold requirement.
Further, the cleaning picture impurity in the step S1 uses manual cleaning, using medical related personnel to aobvious
Micro- picture carries out preliminary screening.5 medical professionals are allowed independently to carry out screening identification, it is ensured that sample database is accurate, more has
Conducive to the raising of discrimination.
Further, it is that there are one worm's ovums in the picture visual field, and have it that the picture in the step S1, which chooses condition,
He interferes impurity, only need to carry out manually marking worm's ovum position and type.
Further, when new samples library discrimination is less than 95%, the artificial mark picture in step S1 is rejected, is transported again
Row step S1~S4, exports result again.
Further, the neural network algorithm used in the step S2 is convolutional neural networks.Usually, the base of CNN
This structure includes two layers, and one is characterized extract layer, and the input of each neuron is connected with the local acceptance region of preceding layer, and carries
Take the feature of the part.After the local feature is extracted, its position relationship between other feature is also decided therewith；
The second is Feature Mapping layer, each computation layer of network is made of multiple Feature Mappings, and each Feature Mapping is a plane, is put down
The weights of all neurons are equal on face.Feature Mapping structure is using the small sigmoid functions of influence function core as convolution net
The activation primitive of network so that Feature Mapping has shift invariant.Further, since the shared power of neuron on a mapping face
Value, thus reduce the number of network freedom parameter.Each convolutional layer followed by one in convolutional neural networks is used for
The computation layer of local average and second extraction, this distinctive structure of feature extraction twice is asked to reduce feature resolution.
This method main advantage has：1. this method establishes large sample number under conditions of setting meets clinical practice relatively
According to library, while invalid value option is introduced, using artificial intelligent depth learning model, learnt automatically from the image that sample database marks
And characteristic information is extracted, so as to identify parasite egg type, quantity, position in the image shot in clinical examination practice
It sets and confidence level.2. invalid value will be automatically recognized as with the incoherent picture of training image, meet what actual use kind to be identified
Picture diversity is higher.3. establishing diversity sample database, picture is acquired at different conditions, is all adopted as long as meeting basic demand
It takes, to improve anti-interference ability, it is ensured that all kinds of interference that encounter when clinically applying occur.4. adjustable accommodating reliability area is arranged
Between (0% -100%), meet different classes of personnel and use.
Description of the drawings
Attached drawing described herein is used for providing further understanding the embodiment of the present invention, constitutes one of the application
Point, do not constitute the restriction to the embodiment of the present invention.In the accompanying drawings：
Fig. 1 is the original displaing micro picture figure of the present invention.
Fig. 2, which is the present invention, need to manually mark original displaing micro picture figure.
Fig. 3 is displaing micro picture figure after the artificial mark of the present invention.
Fig. 4 is the method for the present invention flow chart.
Specific implementation mode
To make the objectives, technical solutions, and advantages of the present invention clearer, with reference to embodiment and attached drawing, to this
Invention is described in further detail, and exemplary embodiment of the invention and its explanation are only used for explaining the present invention, do not make
For limitation of the invention.
Embodiment
As shown in figures 1-4, the present invention is a kind of recognition methods the method packet of the parasite egg based on deep learning
Include following steps：S1：Parasite egg picture is zoomed into displaing micro picture, displaing micro picture is identified, and clears up and does not include
The picture of worm's ovum chooses the picture after cleaning after cleaning does not include the picture of worm's ovum；S2：By the micrograph of selection
Piece establishes sample database, marks parasite egg type and position in sample database displaing micro picture, utilizes artificial intelligence depth
Model is practised to learn automatically from displaing micro picture sample database and extract characteristic information；S3：Utilize the parasitism marked in step S2
Worm's ovum type extracts the characteristic value of parasite egg, establishes property data base, and test samples library identifies in property data base
Rate, and recognition result is exported, and according to discrimination as a result, rejecting fails the sample database picture of identification, formation new samples library.
S4：By the new samples library in step S3, using neural network algorithm, optimization extraction characteristic value forms new feature Value Data library, profit
With new feature Value Data library, new samples library discrimination is examined, inspection result is exported.Currently, the research of parasite egg identification
There are many kinds of methods, are all much the modes of laboratory operation, this kind of laboratory research is often all based on more idealistic
It is carried out under state, experimental method difficulty is competent at actual detection needs, especially in the case where impurity is more, it is difficult to obtain steady
Fixed recognition result, while recognition efficiency is not high.When hospital is studied or carries out clinical analysis of diagnosis, need to parasitism
Worm worm's ovum carries out Classification and Identification, and under existing methods, the recognition methods recognition efficiency of use is low, is again using full manual method
What many hospitals can not undertake, therefore present specification reduces professional to the greatest extent under the premise of effectively improving recognition methods
Utilization, the requirement of discrimination can either be met, and cost is reduced for hospital.
First, worm's ovum picture is amplified under the microscope, amplification becomes displaing micro picture, and enlargement ratio at least needs
400 times because 40X10 (more than) microscope under displaing micro picture, worm's ovum feature is apparent in the visual field, if less than 400 times
Words, amplification effect are bad, it is easy to occur many impurity and extra worm's ovum in the visual field, be unfavorable for picture analyzing.After choosing
I.e. clear up after displaing micro picture establish sample database, in sample database using convolutional neural networks algorithm carry out characteristic value extraction with
Then optimization forms a characteristic value data library, when threshold value of the result discrimination of identification higher than a setting, here
Preferably 95%, discrimination is excessively high, and the consumption of energy consumption can be very big, if discrimination is less than 90%, is also unfavorable for the reality of hospital
It uses, what preferably discrimination 95% can be best takes into account energy consumption and actual use.When discrimination is higher than threshold value, then successfully output is tied
If image after identification while exporting primary result, is established a new samples picture library by fruit preset less than this, secondary logical
It crosses convolution neural network algorithm and carries out characteristics extraction with after optimization, form a new feature Value Data library, export result.Work as threshold
It when value is consistently lower than threshold value, repeats the above steps always, until discrimination reaches threshold requirement.
Specifically, the acquisition of displaing micro picture is carried out first, the clinical front-line expert independent manual in 5, source of displaing micro picture
After identifying and clearing up impurity picture, satisfactory worm's ovum displaing micro picture totally 11263.Wherein certain disease prevention and control center carries
For 825, certain people's hospital laboratory provides 4695, certain Agriculture and Animal Husbandry College provides 5743.The selection condition of displaing micro picture is：
Under 400 times of (40 × 10) mirrors, there is worm's ovum in the visual field, then carries out artificial mark worm's ovum.See Fig. 1, Fig. 2, Fig. 3.Obtain characteristic value
Flow establishes sample database with the displaing micro picture of selection, and training group sample database builds deep learning model, introduces machine learning and calculates
Method is learnt from sample database image using artificial intelligent depth learning model (convolutional neural networks CNN) and is extracted feature automatically
Information is established to analyze the parasite egg type described in sample database picture based on the model rule under the conditions of big data
Then, detailed process is shown in Fig. 4.Test group picture is used to examine the validity of recognition methods.Condition A：400 times of (40 × 10) micrographs
Piece has worm's ovum under the visual field.After carrying out 2 wheel training, if the discrimination to new samples library is less than 95%, sample should be re-established
The samples pictures that not can recognize that are rejected in library.
For sample database original image using collecting 11263 altogether, wherein training group 10118, which is opened, (rejects 118 in training process
, it is practical finally to use 10000), test group 1145 opens.It is opened in addition, invalid value picture 50 is added in test group.For the first time with 100
Picture establishes sample database after opening screening, is trained twice；Sample database is established with picture after 1000 screenings for the second time, carries out two
Secondary training；Picture establishes sample database after 10000 screenings of third time, is trained twice.Specific Evaluated effect is shown in Table 1.
1 total evaluation effect table of table
Note：Accuracy rate：The ratio between sample number and total number of samples for correctly classifying.
F1-score：It is the harmonic-mean of accurate rate and recall rate for certain classification, is herein F1- of all categories
The average of score.
Accurate rate：The sample number for the category that be correctly predicted for certain classification be and the total number of samples for being predicted as the category
The ratio between, it is herein the average of accurate rate of all categories.
Recall rate：To be correctly predicted for certain classification the be sample number of the category and the ratio between the total number of samples of the category,
It is the average of recall rate of all categories herein.
Every time after training, 80% picture in random inspection original sample library is examined with the characteristic value data library of foundation.
Before being trained into next round, after the picture for rejecting disqualified upon inspection, new eligible picture is supplemented.It is evaluated and tested data and is shown in Table 2.
2 sample database recruitment evaluation table of table
By years development, oneself warp achieves more achievement in many fields to deep learning, oneself is through forming in terms of very much
The technology and methods of comparative maturity.Deep learning is this time applied to parasite egg under microscope for the first time and identifies field, is led to
It crosses and establishes big data sample database, discrimination opposite can meet adjuvant clinical diagnosis and treatment at present.Develop into cell phone application and computer end
Doctor's diagnosis and treatment are assisted in some one lines of pilot hospital inspection.
Above-described specific implementation mode has carried out further the purpose of the present invention, technical solution and advantageous effect
It is described in detail, it should be understood that the foregoing is merely the specific implementation mode of the present invention, is not intended to limit the present invention
Protection domain, all within the spirits and principles of the present invention, any modification, equivalent substitution, improvement and etc. done should all include
Within protection scope of the present invention.
Claims (6)
1. a kind of recognition methods of the parasite egg based on deep learning, which is characterized in that described method includes following steps：
S1：Parasite egg picture is zoomed into displaing micro picture, displaing micro picture is identified, and clears up the figure for not including worm's ovum
Piece chooses the picture after cleaning after cleaning does not include the picture of worm's ovum；
S2：The displaing micro picture of selection is established into sample database, marks parasite egg type and position in sample database displaing micro picture
It sets, learnt automatically from displaing micro picture using artificial intelligence deep learning model and extracts characteristic information；
S3：Body worm's ovum type is posted using marking in step S2, the characteristic value of parasite egg is extracted, establishes characteristic
Library, the test samples library discrimination in property data base, and recognition result is exported, and failed as a result, rejecting according to discrimination
The sample database picture of identification forms new samples library.
S4：By the new samples library in step S3, using neural network algorithm, optimization extraction characteristic value forms new feature Value Data
New samples library discrimination is examined, inspection result will be exported using new feature Value Data library in library.
2. a kind of recognition methods of parasite egg based on deep learning according to claim 1, which is characterized in that institute
Stating the cleaning in step S1, the picture comprising worm's ovum does not use manual cleaning, is carried out just to displaing micro picture using medical related personnel
Step screening.
3. a kind of recognition methods of parasite egg based on deep learning according to claim 1, which is characterized in that institute
It is to have multiple worm's ovums in the picture visual field to state the picture in step S1 and choose condition, carries out artificial mark worm's ovum position and type.
4. a kind of recognition methods of parasite egg based on deep learning according to claim 1, which is characterized in that when
When new samples library discrimination is less than 95%, the artificial mark picture in step S1 is rejected, reruns step S1~S4, it is again defeated
Go out result.
5. a kind of recognition methods of parasite egg based on deep learning according to claim 1, which is characterized in that institute
The neural network algorithm used in step S2 is stated as convolutional neural networks algorithm.
6. a kind of recognition methods of parasite egg based on deep learning according to claim 1, which is characterized in that institute
It states displaing micro picture in step S1 and at least amplifies 400 times.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810688609.0A CN108805101A (en) | 2018-06-28 | 2018-06-28 | A kind of recognition methods of the parasite egg based on deep learning |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810688609.0A CN108805101A (en) | 2018-06-28 | 2018-06-28 | A kind of recognition methods of the parasite egg based on deep learning |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN108805101A true CN108805101A (en) | 2018-11-13 |
Family
ID=64071105
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810688609.0A Pending CN108805101A (en) | 2018-06-28 | 2018-06-28 | A kind of recognition methods of the parasite egg based on deep learning |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108805101A (en) |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20040253660A1 (en) * | 2003-06-12 | 2004-12-16 | Giles Scientific, Inc. | Automated microbiological testing apparatus and method |
| US20100086948A1 (en) * | 2008-10-06 | 2010-04-08 | Somalogic, Inc. | Ovarian Cancer Biomarkers and Uses Thereof |
| CN103093235A (en) * | 2012-12-30 | 2013-05-08 | 北京工业大学 | Handwriting digital recognition method based on improved distance core principal component analysis |
| CN103279760A (en) * | 2013-04-09 | 2013-09-04 | 杭州富光科技有限公司 | Real-time classifying method of plant quarantine larvae |
| CN104239858A (en) * | 2014-09-05 | 2014-12-24 | 华为技术有限公司 | Method and device for verifying facial features |
| CN104392240A (en) * | 2014-10-28 | 2015-03-04 | 中国疾病预防控制中心寄生虫病预防控制所 | Parasite egg identification method based on multi-feature fusion |
| CN104700112A (en) * | 2015-02-02 | 2015-06-10 | 电子科技大学 | Method for detecting parasite eggs in excrement based on morphological characteristics |
| CN104992142A (en) * | 2015-06-03 | 2015-10-21 | 江苏大学 | Pedestrian recognition method based on combination of depth learning and property learning |
-
2018
- 2018-06-28 CN CN201810688609.0A patent/CN108805101A/en active Pending
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20040253660A1 (en) * | 2003-06-12 | 2004-12-16 | Giles Scientific, Inc. | Automated microbiological testing apparatus and method |
| US20100086948A1 (en) * | 2008-10-06 | 2010-04-08 | Somalogic, Inc. | Ovarian Cancer Biomarkers and Uses Thereof |
| CN103093235A (en) * | 2012-12-30 | 2013-05-08 | 北京工业大学 | Handwriting digital recognition method based on improved distance core principal component analysis |
| CN103279760A (en) * | 2013-04-09 | 2013-09-04 | 杭州富光科技有限公司 | Real-time classifying method of plant quarantine larvae |
| CN104239858A (en) * | 2014-09-05 | 2014-12-24 | 华为技术有限公司 | Method and device for verifying facial features |
| CN104392240A (en) * | 2014-10-28 | 2015-03-04 | 中国疾病预防控制中心寄生虫病预防控制所 | Parasite egg identification method based on multi-feature fusion |
| CN104700112A (en) * | 2015-02-02 | 2015-06-10 | 电子科技大学 | Method for detecting parasite eggs in excrement based on morphological characteristics |
| CN104992142A (en) * | 2015-06-03 | 2015-10-21 | 江苏大学 | Pedestrian recognition method based on combination of depth learning and property learning |
Non-Patent Citations (3)
| Title |
|---|
| 朱道明等: "《建筑安防技术》", 31 January 2013 * |
| 杨铁军: "《产业专利分析报告(第33册)--智能识别》", 30 June 2015 * |
| 韩瑞珍: "基于机器视觉的农田害虫快速检测与识别研究", 《中国优秀博硕士学位论文全文数据库(博士)》 * |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104992142B (en) | A kind of pedestrian recognition method being combined based on deep learning and attribute study | |
| Khan et al. | Deep learning for apple diseases: classification and identification | |
| CN110046631A (en) | System and method for inferring the variation of time-space image automatically | |
| CN106295661A (en) | The plant species identification method of leaf image multiple features fusion and device | |
| CN106980852A (en) | Based on Corner Detection and the medicine identifying system matched and its recognition methods | |
| CN110731773B (en) | Abnormal electrocardiogram screening method based on fusion of global and local depth features of electrocardiogram | |
| CN109684967A (en) | A kind of soybean plant strain stem pod recognition methods based on SSD convolutional network | |
| CN110084150A (en) | A kind of Automated Classification of White Blood Cells method and system based on deep learning | |
| CN113069080B (en) | Difficult airway assessment method and device based on artificial intelligence | |
| CN108921814A (en) | A kind of Citrus Huanglongbing pathogen on-line quick detection system and method based on deep learning | |
| CN110415818A (en) | A kind of intelligent pediatric disease interrogation system and method based on observable illness | |
| CN109636791A (en) | A kind of Citrus Huanglongbing pathogen detection method based on deep learning, apparatus and system | |
| CN115690072A (en) | Chest radiography feature extraction and disease classification method based on multi-mode deep learning | |
| CN110443105A (en) | The immunofluorescence image kenel recognition methods of autoimmunity antibody | |
| Li et al. | FecalNet: Automated detection of visible components in human feces using deep learning | |
| CN106725341A (en) | A kind of enhanced lingual diagnosis system | |
| CN109829901A (en) | A kind of fungal keratitis detection method and system based on convolutional neural networks | |
| CN109190571A (en) | A kind of detection recognition method and its device of grazing sheep feeding typical plant type | |
| Monigari et al. | Plant leaf disease prediction | |
| Huang et al. | Detection of Diseases Using Machine Learning Image Recognition Technology in Artificial Intelligence | |
| Asare et al. | Detection of anaemia using medical images: A comparative study of machine learning algorithms–A systematic literature review | |
| CN105809200A (en) | Biologically-inspired image meaning information autonomous extraction method and device | |
| Murugeswari et al. | Automated sugarcane disease detection using faster RCNN with an android application | |
| Qayyum et al. | Malaria diagnosis with dilated convolutional neural network based image analysis | |
| CN110348326A (en) | The family health care information processing method of the identification of identity-based card and the access of more equipment |
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 |
Application publication date: 20181113 |
|
| RJ01 | Rejection of invention patent application after publication |