CN111652167A - Intelligent evaluation method and system for chromosome karyotype image - Google Patents
Intelligent evaluation method and system for chromosome karyotype image Download PDFInfo
- Publication number
- CN111652167A CN111652167A CN202010515661.3A CN202010515661A CN111652167A CN 111652167 A CN111652167 A CN 111652167A CN 202010515661 A CN202010515661 A CN 202010515661A CN 111652167 A CN111652167 A CN 111652167A
- Authority
- CN
- China
- Prior art keywords
- chromosome
- image data
- image
- data
- network model
- 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
Images
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
- G06V20/698—Matching; Classification
-
- 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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V2201/00—Indexing scheme relating to image or video recognition or understanding
- G06V2201/04—Recognition of patterns in DNA microarrays
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Biomedical Technology (AREA)
- Health & Medical Sciences (AREA)
- General Physics & Mathematics (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Evolutionary Computation (AREA)
- Data Mining & Analysis (AREA)
- Artificial Intelligence (AREA)
- General Engineering & Computer Science (AREA)
- Computational Linguistics (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Biophysics (AREA)
- Multimedia (AREA)
- Image Analysis (AREA)
- Investigating Or Analysing Biological Materials (AREA)
Abstract
The invention provides a chromosome karyotype image intelligent evaluation method and system, and relates to the technical field of image processing.
Description
Technical Field
The invention relates to the technical field of chromosome karyotype analysis, in particular to a chromosome karyotype image intelligent evaluation method and system.
Background
The karyotype analysis is to examine metaphase chromosome, and to analyze, compare, sort and number chromosome by banding technique according to the characteristics of chromosome length, centromere position, long-short arm ratio, and whether chromosome is associated, and to detect and analyze chromosome structure and number variation. Karyotyping of chromosomes can provide important basis for research of cytogenetic classification, genetic relationship between species and chromosome number and structural variation, karyotyping of chromosomes is an indispensable important means for researching chromosome number, structure and morphological abnormality, whether chromosome number is abnormal or not, whether chromosome fragment deletion, insertion, repetition or the like occurs or not can be judged through karyotyping of chromosomes, and karyotyping of chromosomes is a basic method and gold standard for medical genetics research.
With the development of the technology, the chromosome karyotype analysis technology is mature and is widely applied all over the world, in the current clinical practice, the chromosome karyotype analysis almost completely depends on manpower in the process of examining and reading, and the professional requirements and the actual experience requirements of the examining and reading work on workers are extremely high, so that the wide development of the technology in genetics is hindered, meanwhile, the chromosome karyotype analysis needs to be carried out in a medical institution with a certain scale, the time cost and the labor cost of the personnel to be detected are increased, and the technology is extremely inconvenient.
Therefore, there is a need to provide a method and a system for evaluating chromosome karyotype images, which can reduce the workload of cytogenetics experts, enable the cytogenetics experts to concentrate on finding chromosome abnormalities, improve the working efficiency and the working quality, support remote chromosome karyotype analysis, reduce the cost of chromosome karyotype analysis, perform chromosome karyotype analysis in non-central areas, and widely investigate chromosome diseases, thereby bringing remarkable social and economic benefits.
Disclosure of Invention
The invention provides a chromosome karyotype image intelligent evaluation method and system, which are characterized in that a deep neural network is generated by training based on a large number of original chromosome karyotype images and clinical diagnosis reports through an artificial intelligent deep learning technology, the chromosome karyotype images can be automatically segmented and subjected to subsequent chromosome arrangement and counting, and evaluation results are output through a neural network model.
The invention provides an intelligent evaluation method of a chromosome karyotype image, which comprises the following steps:
s1: collecting karyotype image data;
collecting a large amount of random original chromosome karyotype image data through a cloud platform;
s2: labeling the chromosome image data, constructing a data set for weak supervised learning, and labeling the data by adopting a weak supervised learning method;
s3: processing chromosome image data;
s4: augmentation of chromosome image data;
s5: outputting an evaluation result through a neural network model;
further, the method for labeling the chromosome image by adopting the weak supervised learning method comprises the following steps:
s21: training a depth network model through the marked image data;
firstly, manually labeling a large amount of acquired chromosome image data, constructing a data set D through labeled data and labeled data, and training a depth network model through the labeled data to obtain a first mapping relation F' (x) ═ y, wherein x represents the labeled image data, and y represents the output labeling result;
s22: marking the unmarked image through a depth network model;
labeling the residual unlabeled image data in the data set according to the obtained first mapping relation, and outputting a labeling result and a confidence coefficient p;
s23: optimizing the data set according to the confidence coefficient;
correcting the labeled result data through the confidence coefficient p to obtain a new data set D ', and training the depth network model again according to the obtained data set D' to obtain a second mapping relation F (x) ═ y;
s24: and labeling the image to be labeled through the second mapping relation, outputting confidence coefficient, and outputting a new data set D' after optimization.
And repeating the processes of S21-S24 to obtain the final accurately labeled model mapping relationship f (x) -y.
Further, the processing of the chromosome image comprises automatic gray balance, background whitening and picture removal;
background whitening processing is carried out on the chromosome image data subjected to background gray balance through a background semantic segmentation algorithm;
the picture cleaning automatically identifies and removes blood cells and cell debris and other foreign parts in the chromosome image data through a deep neural network.
Further, the augmentation of the chromosome image data includes: carrying out data augmentation by operations of projection transformation, contrast normalization and sharpening;
the projection transformation is added with controllable coefficients, and matrix multiplication is carried out on the original image coordinates by adopting a 3 x 3 transformation matrix to realize random projection transformation;
the contrast ratio normalization automatically adjusts the contrast ratio of the chromosome image according to different dyeing conditions, and improves the visibility of the display band;
the sharpening adopts a sharpening algorithm to make the chromosome image clearer.
Further, the chromosome karyotype image evaluation is analyzed by adopting a ResNet residual network model, the structure of the residual network model is removed of the large-amplitude pooling operation started by ResNet, meanwhile, the second pooling operation in ResNet is removed, a characteristic diagram with larger chromosome image data is obtained, meanwhile, a non-local attention mechanism is added in the characteristic fusion process, the characteristic diagram after attention weighting is obtained, and then the attention of the network model to the overall distribution is strengthened through the inter-channel self-attention and the spatial self-attention, so that the characteristic diagram Mc of the channel attention and the characteristic diagram Ms of the spatial attention are respectively obtained.
Based on the method, the invention also provides an intelligent evaluation system for the chromosome karyotype image, which comprises a chromosome karyotype image cloud platform, a distributed computing server, a database array and an online scoring system, wherein the chromosome karyotype image cloud platform is remotely connected with the distributed computing server, the online scoring system is remotely connected with the distributed computing server through an interface, and the database array is wirelessly connected with the online scoring system;
the chromosome karyotype image cloud platform is used for collecting chromosome image data on line, providing an on-line data annotation tool for data annotation, and storing the annotated image data into a database;
the distributed computing server performs data expansion on the uploaded chromosome image data to be scored through the built weak supervision learning network model and the image processing module, trains the weak supervision learning network model and the chromosome evaluation depth network model, and receives the chromosome image data uploaded by the online scoring system for scoring;
the online scoring system provides a visual operation platform, a user uploads a chromosome image to be scored online, and calls an algorithm stored in the distributed computing server through an interface to compute and feed back a scoring result, and the online scoring system can upload four chromosome image data simultaneously to perform online scoring.
Furthermore, the evaluation system adopts a distributed micro-service architecture, and a scoring algorithm is called through interfaces of all servers to realize cross-cloud computing.
The invention has the advantages that the invention distinguishes the prior art:
1. the method comprises the steps of constructing a data set for weak supervised learning through artificially labeled image data and unlabeled image data, training a model mapping relationship through the image data labeled in the data set, obtaining a new data set, and carrying out multiple times of optimization through confidence level according to manual work to obtain a model mapping relationship with accurate labeling, so that a network model is concentrated in uncertain image data, and the efficiency and accuracy of data labeling are improved.
2. A projection transformation mode is adopted in the process of amplifying chromosome image data, the projection transformation increases a controllable coefficient on the basis of affine transformation to realize random projection transformation, and the change of a focal plane of a microscope is simulated to enhance the capability of a network to adapt to different focal planes.
3. In the scoring network model, two pooling operations are removed based on a ResNet residual network structure, so that an output feature map is expanded by four times, more details in image data are reserved, meanwhile, a non-local attention mechanism is added to strengthen the fusion of global semantic features, and a channel attention mechanism and a space attention mechanism are utilized to strengthen the attention of the network model to feature distribution and strengthen the learning capability of the network model.
4. The system adopts a cloud platform to collect, label and store chromosome image data, adopts a distributed micro-service framework to perform cross-cloud image analysis, and improves the efficiency of the system.
Drawings
FIG. 1 is a flow chart of the intelligent evaluation method of chromosome karyotype images according to the present invention;
FIG. 2 is a schematic diagram of the system architecture of the present invention;
FIG. 3 is a flow chart of the model construction of the weakly supervised learning method of the present invention;
FIG. 4 is a schematic diagram of the image processing process of the present invention;
Detailed Description
In the following description, technical solutions in the embodiments of the present invention are clearly and completely described, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
The embodiment one of the invention provides an intelligent evaluation method of a chromosome karyotype image, as shown in figure one, the method comprises the steps of S1-S5, and the specific contents of the steps are as follows:
s1: collecting chromosome karyotype image data;
according to the embodiment, 83271 chromosome images are collected from various medical institutions, research institutions, medical genetics research centers and laboratories all over the country through the network and are used for building and training a network model, and the comprehensiveness and randomness of chromosome image data are guaranteed.
S2: constructing and training a weak supervised learning network model, and carrying out data annotation by adopting a weak supervised learning method;
as shown in fig. 3, the method steps are as follows:
s21: constructing data set of weak supervised learning and training network model
Through setting an online questionnaire, inviting an expert team to grade the acquired random chromosome image, classifying and sorting the labeled image data to generate image-labeled data, acquiring a labeled standard data set, and forming a data set D by the standard data set and data which are not subjected to manual labeling, wherein the data set is in the form as follows:
D＝{(x1，y1)...(xm，ym)，xm+1，...，xl}
wherein X is image data, X1-XmRepresenting artificially labelled chromosome image data, Xm+1-XlRepresenting chromosome image data which is not subjected to manual labeling, and training m labeled data samples to obtain a first model mapping relation F' (x) ═ y;
s22: mapping and labeling the remaining image data in the data set by using the first model, and simultaneously outputting the labeled confidence coefficient, wherein the obtained output result is as follows:
{(xm，y′m，pm)，...，(xl，y′l，pl)}
wherein p represents the confidence of the network model annotation image;
s23: performing manual examination according to the confidence coefficient, and correcting the labeled image data;
according to the confidence coefficient of the image labeled by the network model, when the confidence coefficient is less than 0.8, judging that the labeled image data is inaccurate, manually labeling the image data again, and obtaining a new data set D ', wherein D' is as follows:
{(x1，y1)...(xm，ym)，(x″m+1，y″m+1)，...，(x″n，y″n) Retraining the acquired chromosome image data in the dataset to obtain a second mapping relationship, F "(x) y;
s24: mapping image data to be annotated in the annotation statement data set by using a second model, outputting confidence coefficient, and optimizing again according to the obtained execution degree to obtain a new data set;
repeating the processes of S21-S24 until the confidence of all the output image data labeled by the network model is greater than 0.8, and obtaining the nth mapping relationship f (x) -y as the final mapping relationship labeled accurately.
S3: carrying out image processing on the obtained labeled chromosome image;
as shown in fig. 4, the labeled chromosome image is subjected to automatic gray balance processing, the background of the chromosome image is whitened by automatically cutting off the background through a background semantic segmentation algorithm, and finally, blood cells, cell fragments and other impurity parts are automatically identified and removed by using a deep neural network;
s4: amplifying the obtained chromosome image data after image processing by a special method;
the obtained chromosome image data is obtained by shooting through a high-definition microscope, each chromosome image data has different microscope focal plane changes, and projection transformation, contrast normalization and image sharpening are adopted to amplify the chromosome image data in order to enable the training data of the scoring network model to be more practical;
the projection transformation adds a controllable coefficient on the basis of affine transformation, the controllable coefficient is a randomly generated 3 x 3 transformation matrix, the transformation matrix is multiplied by the original image coordinates to obtain image data after random projection transformation, and the transformation formula is as follows:
and the contrast is normalized, the contrast is automatically adjusted according to different conditions of chromosomes in the image data, and normalization calculation is carried out, wherein the parameters of the normalization calculation in the present example are as follows: mean ═ 0.485, 0.456, 0.406], std ═ 0.229, 0.224, 0.225;
the image sharpening adopts a USM sharpening algorithm to sharpen chromosome image data, and the sharpening parameters in the embodiment are as follows: the amount is 0.1; the radius is 1.5; threshold value: 0.
s5: constructing and training a scoring network model;
in the construction of the scoring network model, the ResNet residual error network structure is adjusted, so that the method is better applied to chromosome karyotype analysis;
firstly, removing the operation of large-amplitude pooling started by a ResNet residual network model, simultaneously removing the second of 3 pooling operations of the ResNet residual network model, outputting the input n x n image data to obtain a n/8 x n/8-sized feature map, amplifying the feature map by four times, reserving more details in the image data, and avoiding detail loss caused by small chromosome size;
secondly, a non-local attention mechanism is added into the neural network model, so that the global semantic feature fusion is enhanced, and the feature expression capability in the network model is improved;
obtaining the non-local attention consistent with the size of the feature map by using a non-local attention calculation formula on the feature map output by the residual error network model convolutional layer, wherein the non-local attention calculation formula is as follows:
wherein g (x) is a network model mapping function, f (x) is a similarity measurement function, c (x) is a weight, and the obtained attention is calculated by the following weighting formula to obtain a weighted feature map, wherein the weighting formula is as follows:
zi＝Wz·yi+xi
wherein Wz，xi；
Finally, a channel self-attention mechanism and a space self-attention mechanism are adopted to enhance the attention of the network model to the overall distribution of the characteristics;
the characteristic diagram calculation formula of the channel attention mechanism is as follows:
wherein F is a characteristic diagram output by a ResNet residual error network model convolution layer, the characteristic diagram results of MaxBooling and Global Avg Pooling of the obtained characteristic diagrams are respectively sent to a three-layer perceptron MLP, the output results are added and then sent to a ReLU activation function, and a characteristic diagram Mc of channel attention is obtained;
and inputting the obtained feature map Mc into a spatial attention module for calculation to obtain a final feature map Ms applying a spatial attention mechanism.
The calculation formula of the feature map of the spatial self-attention mechanism is as follows:
and finally, performing feature fusion on the obtained feature map Mc and the feature map Ms, and then performing linear conversion to obtain a final scoring result for outputting.
The embodiment two of the invention provides an intelligent evaluation system for a chromosome karyotype image, which comprises a chromosome karyotype image cloud platform, a distributed computing server, a database array and an online evaluation system, wherein the chromosome karyotype image cloud platform is remotely connected with the distributed computing server, the online evaluation system is remotely connected with the distributed computing server through an interface, the database array is wirelessly connected with the online evaluation system, the evaluation system adopts a distributed micro-service architecture, and a evaluation algorithm is called through the interface of each server to realize cross-cloud computing;
the chromosome karyotype image cloud platform is used for collecting chromosome image data on line, providing an on-line data annotation tool for data annotation, and storing the annotated image data into a database;
the distributed computing server performs data expansion on the uploaded chromosome image data to be scored through the built weak supervision learning network model and the image processing module, trains the weak supervision learning network model and the chromosome evaluation depth network model, and receives the chromosome image data uploaded by the online scoring system for scoring;
the online scoring system provides a visual operation platform, a user uploads a chromosome image to be scored online, and calls an algorithm stored in the distributed computing server through an interface to compute and feed back a scoring result, and the online scoring system can upload four chromosome image data simultaneously to perform online scoring.
The invention is not limited to the foregoing embodiments. The invention extends to any novel feature or any novel combination of features disclosed in this specification and any novel method or process steps or any novel combination of features disclosed.
Claims (7)
1. An intelligent evaluation method for a chromosome karyotype image, the method comprising:
s1: collecting karyotype image data;
s2: labeling the chromosome image data, constructing a data set for weak supervised learning, and labeling the data by adopting a weak supervised learning method;
s3: processing chromosome image data;
s4: augmentation of chromosome image data;
s5: and outputting an evaluation result through a neural network model.
2. The intelligent evaluation method for karyotype images according to claim 1, wherein the weakly supervised learning method in step S2 includes the steps of:
s21: training a depth network model through the marked image data;
firstly, manually labeling a large amount of acquired chromosome image data, constructing a data set D through labeled data and labeled data, and training a depth network model through the labeled data to obtain a first mapping relation F' (x) ═ y, wherein x represents the labeled image data, and y represents the output labeling result;
s22: marking the unmarked image through a depth network model;
labeling the residual unlabeled image data in the data set according to the obtained first mapping relation, and outputting a labeling result and a confidence coefficient p;
s23: optimizing the data set according to the confidence coefficient;
correcting the labeled result data through the confidence coefficient p to obtain a new data set D ', and training the depth network model again according to the obtained data set D' to obtain a second mapping relation F (x) ═ y;
s24: and labeling the image to be labeled through the second mapping relation, outputting confidence, outputting a new data set D after optimization, and repeating the process to obtain the model mapping relation F (x) y with accurate labeling.
3. The intelligent evaluation method for the chromosome karyotype image according to claim 1, wherein the processing of the chromosome image includes automatic gray balancing, background whitening, picture removal;
background whitening processing is carried out on the chromosome image data subjected to background gray balance through a background semantic segmentation algorithm;
the picture cleaning automatically identifies and removes blood cells and cell debris and other foreign parts in the chromosome image data through a deep neural network.
4. The intelligent evaluation method for chromosome karyotype images according to claim 1, wherein the augmentation of chromosome image data includes: carrying out data augmentation by operations of projection transformation, contrast normalization and sharpening;
and the projection transformation is added with controllable coefficients, and the original image coordinates are subjected to matrix multiplication by using a 3 x 3 transformation matrix, so that random projection transformation is realized.
5. The intelligent evaluation method for the karyotype image according to claim 1, wherein in S6, the karyotype image data is analyzed using a ResNet residual network model;
the residual error network model structure removes the large-amplitude pooling operation started by ResNet, removes the second pooling operation in ResNet, obtains a characteristic diagram with larger chromosome image data, adds a non-local attention mechanism in the characteristic fusion process, obtains a characteristic diagram after attention weighting, and respectively obtains a characteristic diagram Mc of channel attention and a characteristic diagram Ms of space attention by enhancing the attention of the network model to the overall distribution through the inter-channel self-attention and the space self-attention.
6. The intelligent evaluation system for the chromosome karyotype image is characterized by comprising a chromosome karyotype image cloud platform, a distributed computing server, a database array and an online scoring system, wherein the chromosome karyotype image cloud platform is remotely connected with the distributed computing server, the online scoring system is remotely connected with the distributed computing server through an interface, and the database array is wirelessly connected with the online scoring system;
the chromosome karyotype image cloud platform is used for collecting chromosome image data on line, providing an on-line data annotation tool for data annotation, and storing the annotated image data into a database;
the distributed computing server performs data expansion on the uploaded chromosome image data to be scored through the built weak supervision learning network model and the image processing module, trains the weak supervision learning network model and the chromosome evaluation depth network model, and receives the chromosome image data uploaded by the online scoring system for scoring;
the online scoring system provides a visual operation platform, a user uploads a chromosome image to be scored online, and calls an algorithm stored in the distributed computing server through an interface to compute and feed back a scoring result, and the online scoring system can upload four chromosome image data simultaneously to perform online scoring.
7. The intelligent evaluation system for the chromosome karyotype images according to claim 6, wherein the evaluation system adopts a distributed micro-service architecture, and a scoring algorithm is invoked through an interface of each server to realize cross-cloud computing.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010515661.3A CN111652167A (en) | 2020-06-09 | 2020-06-09 | Intelligent evaluation method and system for chromosome karyotype image |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010515661.3A CN111652167A (en) | 2020-06-09 | 2020-06-09 | Intelligent evaluation method and system for chromosome karyotype image |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111652167A true CN111652167A (en) | 2020-09-11 |
Family
ID=72349393
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010515661.3A Pending CN111652167A (en) | 2020-06-09 | 2020-06-09 | Intelligent evaluation method and system for chromosome karyotype image |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111652167A (en) |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112711983A (en) * | 2020-12-08 | 2021-04-27 | 湖南自兴智慧医疗科技有限公司 | Nuclear pattern analysis system, method, electronic device and readable storage medium |
| CN112733873A (en) * | 2020-09-23 | 2021-04-30 | 浙江大学山东工业技术研究院 | Chromosome karyotype graph classification method and device based on deep learning |
| CN112785566A (en) * | 2021-01-15 | 2021-05-11 | 湖南自兴智慧医疗科技有限公司 | Chromosome metaphase image scoring method and device, electronic equipment and storage medium |
| CN113920069A (en) * | 2021-09-26 | 2022-01-11 | 广州达安临床检验中心有限公司 | Construction method, construction device, equipment and storage medium of chromosome karyotype analysis simulation data set |
| CN114170218A (en) * | 2021-12-16 | 2022-03-11 | 易构智能科技(广州)有限公司 | Chromosome image instance label generation method and system |
| CN115082474A (en) * | 2022-08-22 | 2022-09-20 | 湖南自兴智慧医疗科技有限公司 | Chromosome segmentation method and device based on homologous same-class chromosome information |
| CN117275030A (en) * | 2023-09-27 | 2023-12-22 | 自然资源部地图技术审查中心 | Method and device for auditing map |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107609503A (en) * | 2017-09-05 | 2018-01-19 | 刘宇红 | Intelligent cancerous tumor cell identifying system and method, cloud platform, server, computer |
| CN108985334A (en) * | 2018-06-15 | 2018-12-11 | 广州深域信息科技有限公司 | The generic object detection system and method for Active Learning are improved based on self-supervisory process |
| CN109034190A (en) * | 2018-06-15 | 2018-12-18 | 广州深域信息科技有限公司 | A kind of object detecting system and method for the active sample excavation of dynamic select strategy |
| CN109063748A (en) * | 2018-07-16 | 2018-12-21 | 重庆大学 | Object detection method based on data enhancing |
| CN110265087A (en) * | 2018-10-05 | 2019-09-20 | 中国医药大学附设医院 | Chromosome abnormality detection model, its detection system and chromosome abnormality detection method |
| CN110796143A (en) * | 2019-10-31 | 2020-02-14 | 天津大学 | Scene text recognition method based on man-machine cooperation |
| US20200111212A1 (en) * | 2018-10-05 | 2020-04-09 | China Medical University Hospital | Chromosome Abnormality Detecting Model, Detecting System Thereof, And Method For Detecting Chromosome Abnormality |
| CN111104898A (en) * | 2019-12-18 | 2020-05-05 | 武汉大学 | Image scene classification method and device based on target semantics and attention mechanism |
-
2020
- 2020-06-09 CN CN202010515661.3A patent/CN111652167A/en active Pending
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107609503A (en) * | 2017-09-05 | 2018-01-19 | 刘宇红 | Intelligent cancerous tumor cell identifying system and method, cloud platform, server, computer |
| CN108985334A (en) * | 2018-06-15 | 2018-12-11 | 广州深域信息科技有限公司 | The generic object detection system and method for Active Learning are improved based on self-supervisory process |
| CN109034190A (en) * | 2018-06-15 | 2018-12-18 | 广州深域信息科技有限公司 | A kind of object detecting system and method for the active sample excavation of dynamic select strategy |
| CN109063748A (en) * | 2018-07-16 | 2018-12-21 | 重庆大学 | Object detection method based on data enhancing |
| CN110265087A (en) * | 2018-10-05 | 2019-09-20 | 中国医药大学附设医院 | Chromosome abnormality detection model, its detection system and chromosome abnormality detection method |
| US20200111212A1 (en) * | 2018-10-05 | 2020-04-09 | China Medical University Hospital | Chromosome Abnormality Detecting Model, Detecting System Thereof, And Method For Detecting Chromosome Abnormality |
| CN110796143A (en) * | 2019-10-31 | 2020-02-14 | 天津大学 | Scene text recognition method based on man-machine cooperation |
| CN111104898A (en) * | 2019-12-18 | 2020-05-05 | 武汉大学 | Image scene classification method and device based on target semantics and attention mechanism |
Non-Patent Citations (3)
| Title |
|---|
| FISHER YU 等: "Dilated Residual Networks", 《2017 IEEE CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION》 * |
| XIAOLONG WANG 等: "Non-local Neural Networks", 《2018 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION》 * |
| 李康 等: "基于卷积神经网络和几何优化的统计染色体核型分析方法", 《南京大学学报（自然科学）》 * |
Cited By (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112733873A (en) * | 2020-09-23 | 2021-04-30 | 浙江大学山东工业技术研究院 | Chromosome karyotype graph classification method and device based on deep learning |
| CN112711983A (en) * | 2020-12-08 | 2021-04-27 | 湖南自兴智慧医疗科技有限公司 | Nuclear pattern analysis system, method, electronic device and readable storage medium |
| CN112785566A (en) * | 2021-01-15 | 2021-05-11 | 湖南自兴智慧医疗科技有限公司 | Chromosome metaphase image scoring method and device, electronic equipment and storage medium |
| CN112785566B (en) * | 2021-01-15 | 2024-01-19 | 湖南自兴智慧医疗科技有限公司 | Metaphase image scoring method, metaphase image scoring device, electronic equipment and storage medium |
| CN113920069A (en) * | 2021-09-26 | 2022-01-11 | 广州达安临床检验中心有限公司 | Construction method, construction device, equipment and storage medium of chromosome karyotype analysis simulation data set |
| CN113920069B (en) * | 2021-09-26 | 2022-07-08 | 广州达安临床检验中心有限公司 | Construction method, construction device, equipment and storage medium of chromosome karyotype analysis simulation data set |
| CN114170218A (en) * | 2021-12-16 | 2022-03-11 | 易构智能科技(广州)有限公司 | Chromosome image instance label generation method and system |
| CN115082474A (en) * | 2022-08-22 | 2022-09-20 | 湖南自兴智慧医疗科技有限公司 | Chromosome segmentation method and device based on homologous same-class chromosome information |
| CN115082474B (en) * | 2022-08-22 | 2023-03-03 | 湖南自兴智慧医疗科技有限公司 | Chromosome segmentation method and device based on homologous same-class chromosome information |
| CN117275030A (en) * | 2023-09-27 | 2023-12-22 | 自然资源部地图技术审查中心 | Method and device for auditing map |
| CN117275030B (en) * | 2023-09-27 | 2024-05-14 | 自然资源部地图技术审查中心 | Method and device for auditing map |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111652167A (en) | Intelligent evaluation method and system for chromosome karyotype image | |
| WO2022160771A1 (en) | Method for classifying hyperspectral images on basis of adaptive multi-scale feature extraction model | |
| CN112288706B (en) | Automatic chromosome karyotype analysis and abnormality detection method | |
| CN110415230B (en) | CT slice image semantic segmentation system and method based on deep learning | |
| CN111462120A (en) | Defect detection method, device, medium and equipment based on semantic segmentation model | |
| CN113256636A (en) | Bottom-up parasite species development stage and image pixel classification method | |
| CN111724355A (en) | Image measuring method for abalone body type parameters | |
| CN110288017B (en) | High-precision cascade target detection method and device based on dynamic structure optimization | |
| CN111476307B (en) | Lithium battery surface defect detection method based on depth field adaptation | |
| CN113139977B (en) | Mouth cavity curve image wisdom tooth segmentation method based on YOLO and U-Net | |
| CN113919443A (en) | Tobacco maturity state probability calculation method based on image analysis | |
| CN112749675A (en) | Potato disease identification method based on convolutional neural network | |
| CN115602337A (en) | Cryptocaryon irritans disease early warning method and system based on machine learning | |
| CN113222149A (en) | Model training method, device, equipment and storage medium | |
| CN111898677A (en) | Plankton automatic detection method based on deep learning | |
| Kang et al. | Semantic segmentation of fruits on multi-sensor fused data in natural orchards | |
| CN114898220A (en) | Intelligent production control method for structural member of overhead working truck | |
| CN113496148A (en) | Multi-source data fusion method and system | |
| CN114387261A (en) | Automatic detection method suitable for railway steel bridge bolt diseases | |
| CN114399763A (en) | Single-sample and small-sample micro-body ancient biogenetic fossil image identification method and system | |
| CN114399108A (en) | Tea garden yield prediction method based on multi-mode information | |
| Patel et al. | Deep Learning-Based Plant Organ Segmentation and Phenotyping of Sorghum Plants Using LiDAR Point Cloud | |
| CN116091828A (en) | Chromosome image interpretable analysis method, device, equipment and storage medium | |
| CN114913523B (en) | Yolox-based multifunctional real-time intelligent plant stomata recognition system | |
| CN114494441B (en) | Grape and picking point synchronous identification and positioning method and device 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 | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20200911 |
|
| WD01 | Invention patent application deemed withdrawn after publication |